Article Text

Extended report
Open-label tofacitinib and double-blind atorvastatin in rheumatoid arthritis patients: a randomised study*
  1. Iain B McInnes1,
  2. Ho-Youn Kim2,
  3. Sang-Heon Lee3,
  4. David Mandel4,
  5. Yeong-Wook Song5,
  6. Carol A Connell6,
  7. Zhen Luo7,
  8. M Julia Brosnan6,
  9. Andrea Zuckerman6,
  10. Samuel H Zwillich6,
  11. John D Bradley6
  1. 1Glasgow Biomedical Research Centre, University of Glasgow, Glasgow, UK
  2. 2Division of Rheumatology, Department of Internal Medicine, Rheumatism Research Center, The Catholic University of Korea, Seoul, Republic of Korea
  3. 3Division of Rheumatology, Konkuk University School of Medicine, Seoul, Republic of Korea
  4. 4Rheumatology and Osteoporosis Practice, Mayfield Village, Ohio, USA
  5. 5Department of Internal Medicine, Division of Rheumatology, Medical Research Center, Seoul National University College of Medicine, Seoul, Republic of Korea
  6. 6Medicines Development, Pfizer Inc, Groton, Connecticut, USA
  7. 7Pfizer Inc, Shanghai, China
  1. Correspondence to Dr Iain B McInnes, Glasgow Biomedical Research Centre, University of Glasgow, 120 University Place, Glasgow, G12 8QQ, UK; iain.mcinnes{at}


Objectives To evaluate the efficacy and safety of atorvastatin versus placebo in modifying lipids in patients with rheumatoid arthritis (RA) receiving the oral Janus kinase inhibitor, tofacitinib.

Methods A randomised, placebo controlled, multicentre phase 2 study, open-label for tofacitinib and blinded for atorvastatin. Patients received tofacitinib 10 mg twice daily for 12 weeks; at week 6, patients were randomly assigned 1:1 to receive oral atorvastatin 10 mg once daily or placebo for 6 weeks. Main outcome measures were lipid moieties, American College of Rheumatology (ACR) response rates, disease activity score in 28 joint counts and safety.

Results 111 patients meeting ACR 1987 RA criteria with active disease were enrolled. Tofacitinib-induced elevation of mean total, low-density lipoprotein (LDL) and high-density lipoprotein-cholesterol, triglycerides and apolipoprotein A-1 concentrations were sustained in placebo recipients to week 12; atorvastatin added at week 6 significantly reduced tofacitinib-associated increases in total and LDL-cholesterol, triglycerides and apolipoprotein B to below week 0 levels. Co-administration of atorvastatin resulted in a significant reduction of LDL-cholesterol versus placebo (primary endpoint; p<0.0001); from week 6 to week 12 the least squares mean reduction was 35.3% with atorvastatin, versus 5.8% increase with placebo. ACR responses were observed with tofacitinib; numerically greater rates were seen with atorvastatin versus placebo. Adverse events were consistent with phase 3 studies.

Conclusions Tofacitinib-associated elevated total and LDL-cholesterol and triglycerides were rapidly and significantly reduced by atorvastatin. Further investigation is required to explore the significance of reductions in RA disease activity in patients receiving tofacitinib and atorvastatin.

( identifier NCT01059864; Pfizer protocol A3921109).

  • Rheumatoid Arthritis
  • Lipids
  • Treatment

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Rheumatoid arthritis (RA) is a chronic inflammatory disease associated with articular inflammation and damage, progressive disability, functional decline and accelerated mortality, mainly due to cardiovascular events.1 ,2 Current treatment comprises conventional disease-modifying antirheumatic drugs (DMARD) followed by biological DMARD, if necessary, to achieve low disease activity or remission. Therapeutics used in RA exhibit limitations in tolerability, access and response duration and magnitude. There remains keen interest in developing novel therapeutic agents.3

The receptor signalling cascades mediating inflammatory pathways directly implicated in RA pathogenesis are becoming clearer and are the focus of research as therapeutic targets.2 ,4–6 The Janus kinase (JAK) family is pivotal in regulating signalling by T, B and natural killer cells, implicated in synovitis and articular damage.2 ,4 Increased understanding of the structural and signalling characteristics of kinases has enabled the development of a range of blocking agents, including small, orally bioavailable molecules.7

Tofacitinib is a novel, oral JAK inhibitor that is being investigated as a targeted immunomodulator and DMARD,7 ,8 which preferentially inhibits signalling by receptors associated with JAK3 and JAK1 over JAK2, and blocks signalling through the common gamma chain-containing receptors for cytokines including interleukin (IL)-2, IL-4, IL-7, IL-9, IL-15 and IL-21;8 ,9 these are integral to lymphocyte function, and inhibition of their signalling may result in immune response modulation.7 In addition, JAK1 inhibition results in attenuation of signalling by pro-inflammatory cytokines, including IL-6 and type I interferon. Encouraging efficacy has been observed in phase 210–13 and phase 314–18 studies thus far.

As in other inflammatory states, patients with active RA may exhibit reduced circulating cholesterol levels that ‘normalise’ on effective therapy.19 Elevated cholesterol and triglyceride levels beyond reference ranges for a given population have been observed after initiation of therapy with biological DMARD; the magnitude appears greater with IL-6 receptor versus tumour necrosis factor inhibition.

Dose-dependent mean increases in total, high-density lipoprotein (HDL) and low-density lipoprotein (LDL) cholesterol of 16–30% were observed in patients with RA receiving tofacitinib.10–13 The mechanism(s) responsible for these changes is unclear and the effect of a 5-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitor remains unknown. Atorvastatin 10 mg daily has been shown to reduce LDL-cholesterol by 26–37%, as well as total cholesterol and apolipoprotein B, in populations with or at risk of cardiovascular disease.20 We studied the magnitude and character of lipid elevations in patients with RA receiving tofacitinib, within a controlled trial setting and controlled lifestyle diet, and herein document the effects of atorvastatin on lipid elevations and RA disease activity.


Design overview

This randomised, 12-week phase 2 study was designed to evaluate the safety and efficacy of atorvastatin versus placebo in modulating LDL-cholesterol in patients with active RA receiving tofacitinib. The study was conducted in 15 centres in the USA and Korea between 11 February and 22 November 2010, in compliance with the Declaration of Helsinki and the International Conference on Harmonisation good clinical practice guidelines. The final protocol, amendments and informed consent documentation were approved by the institutional review boards and independent ethics committees of the investigational centres. All patients provided written informed consent.

Setting and participants

Eligible patients were 18 years of age or older and were diagnosed with RA based on American College of Rheumatology (ACR) 1987 revised criteria.21 Active disease was defined as four or more tender/painful joints and four or more swollen joints and either erythrocyte sedimentation rate (ESR; Westergren method) greater than 28 mm/h or C-reactive protein (CRP) greater than 7 mg/l. Intravenous, intra-articular or intramuscular corticosteroids, intra-articular hyaluronate sodium, biological response modifiers and DMARD (investigational or marketed) were not permitted less than 4 weeks before the first tofacitinib dose (or less than 6 (adalimumab), 8 (influximab, auranofin, aurothioglucose, leflunomide), 10 (golimumab) or 12 weeks (abatacept)). Stable doses of non-steroidal anti-inflammatory drugs, selective cyclooxygenase-2 inhibitors, opioids, acetaminophen and low-dose oral corticosteroids were allowed. Before the first tofacitinib dose, all lipid-lowering agents were stopped for 3 months; other prohibited medications for 4 weeks or five half-lives (whichever was longer). Supplements containing plant sterols/stanols or cholestin were not permitted.

Key exclusion criteria: haemoglobin less than 9.0 g/dl, haematocrit less than 30%, white blood cell count less than 3.0×109/l, absolute neutrophil count less than 1.2×109/l or platelet count less than 100×109/l; estimated glomerular filtration rate 40 ml/min or less (Cockcroft–Gault calculation); aspartate aminotransferase or alanine aminotransferase greater than 1.5× upper limit of normal; fasting triglycerides greater than 4.52 mmol/l and fasting total cholesterol greater than 8.29 mmol/l; history of malignancy except adequately treated or excised non-metastatic basal/squamous cell cancer of the skin or cervical carcinoma in situ; history of herpes zoster, hepatitis B/C or HIV; and evidence of Mycobacterium tuberculosis infection, determined by a negative medical history and physical examination, a chest x-ray negative for active tuberculosis and a negative QuantiFERON-TB Gold In-Tube (Cellestis, Valencia, California, USA) test performed less than 3 months before screening.

Randomisation and interventions

The study was open-label for tofacitinib, and patient-investigator- and sponsor-blinded for atorvastatin. Patients received oral tofacitinib 10 mg twice daily (bid) for 12 weeks; at week 6, patients were randomly assigned 1:1 to oral atorvastatin 10 mg once daily or placebo, using an interactive voice response system (see supplementary text, available online only).

The sample size of 100 (50/group) was determined primarily based on safety database requirements; this achieves a power of over 90% at a type I error rate of 0.025 (one-sided) to establish superiority of atorvastatin/tofacitinib versus placebo/tofacitinib in percentage LDL-cholesterol change from weeks 6 to 12, assuming a treatment difference of 15% and a common SD of 20%.

Outcomes and follow-up

The primary objective was to evaluate the safety and efficacy of 6 weeks’ atorvastatin versus placebo in reducing LDL-cholesterol in patients with RA receiving tofacitinib. The primary endpoint was the percentage change in the LDL-cholesterol level from week 6 (baseline) to week 12. Secondary endpoint lipoprotein parameters included absolute changes in LDL-cholesterol, total cholesterol, HDL-cholesterol, apolipoprotein A-1 and B and triglycerides from weeks 6 to 12; and 12-h fasting lipid profiles at each visit (LipoScience, Raleigh, North Carolina, USA). Secondary endpoints also included ACR response rates and core components, and disease activity score in 28 joints (DAS28)-4(ESR) and response rates (<2.6 and ≤3.2) at weeks 6 and 12.

The incidence and severity of adverse events (AE) and clinical laboratory abnormalities were recorded; vital-sign assessments and physical examinations were performed at each visit. ECG were performed at screening and week 12.

Statistical analysis

The primary endpoint was analysed based on the full analysis set (FAS, a modified intention-to-treat population, defined as patients who received one or more dose of randomised investigational drug) using a linear mixed-effects model with treatment, visit and treatment-by-visit interaction as fixed-effects and baseline LDL-cholesterol level as a covariate. Correlations among measurements on the same patient at different visits were modelled by including a common random effect in the model. All least squares (LS) means were calculated based on this model. The primary analysis assessed superiority of atorvastatin over placebo in LDL-cholesterol reduction (significance level 0.025; 1-sided). Additional lipid profile endpoints, including those from clinical laboratory tests and a specialty laboratory, LipoScience (Raleigh, North Carolina, USA), were analysed similarly as described for the primary endpoint and no correction for multiple comparisons was applied. RA disease activity endpoints were summarised by treatment and visit using descriptive statistics.

The open-label safety analysis set included patients who received one or more dose of open-label tofacitinib, referred to as the safety population (includes patients from weeks 0 to 12 unless otherwise specified). The double-blind safety analysis set was the same as the FAS.



Of 198 patients screened, 111 were enrolled and received tofacitinib for 12 weeks or less. During the double-blind period (weeks 6–12), 50 patients were randomly assigned to additional treatment with atorvastatin (25 each from the USA and Korea) and 47 were randomly assigned to additional placebo (30 from the USA and 17 from Korea); 92 (94.8%) patients completed the study (figure 1). Most patients were women (89.7%); mean age 52 years; 43 patients were Asian: 42 from Korea and one from the USA. Disease duration since diagnosis ranged from 6 weeks to 37 years (mean duration 9.2 years); 12 patients receiving atorvastatin/tofacitinib and 11 patients receiving placebo/tofacitinib did not have CRP greater than 7 mg/l, so qualified based on ESR. Demographic and enrolment characteristics are presented in table 1.

Table 1

Demography and patient characteristics at enrolment

Figure 1

Patient disposition. bid, twice daily; FAS, full analysis set; QD, once daily.

Atorvastatin effectively reduces tofacitinib-induced lipid elevations

Mean LDL-cholesterol levels were similar in the atorvastatin/tofacitinib and placebo/tofacitinib groups at week 0; and increased in both groups by week 6 (figure 2A). This change was characterised by an increased ratio of large LDL/total LDL with a compensatory reduction in small LDL/total LDL (see supplementary table S1, available online only). Co-administration of atorvastatin to patients receiving tofacitinib resulted in a significant percentage reduction of LDL-cholesterol versus placebo (p<0.0001): from weeks 6 to 12 the LS mean reduction was 35.3% with atorvastatin, versus a 5.8% increase with placebo. The LS mean absolute change in LDL-cholesterol from weeks 6 to 12 was also different (p<0.0001) in the atorvastatin versus placebo group (see supplementary table S1, available online only). No further changes in the large LDL/total LDL ratio was observed in either group from weeks 6 to 12 (see supplementary table S1, available online only).

Figure 2

Mean lipoprotein cholesterol, total cholesterol, triglyceride and apolipoprotein levels over time. The horizontal dashed line in panel A indicates the ATPIII optimal target range (<100 mg/dl). bid, twice daily; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

Mean total cholesterol levels were similar at week 0, and increased to week 6 in both groups (figure 2B). The addition of atorvastatin resulted in significant percentage and absolute reduction in total cholesterol by week 12 (−21.1%; absolute −1.36 mmol/l; p<0.0001) versus placebo (3.8%; absolute 0.19 mmol/l). HDL-cholesterol levels rose in both groups from week 0 to week 6. Percentage and absolute changes in HDL-cholesterol levels from weeks 6 to 12 were unaltered and similar in each group after the addition of atorvastatin or placebo (figure 2C and see supplementary table S1, available online only). Ratios of total cholesterol/HDL-cholesterol and LDL-cholesterol/HDL-cholesterol were unchanged from weeks 0 to 6 with tofacitinib, and dropped after the addition of atorvastatin (see supplementary table S1, available online only).

Triglyceride levels were lower by chance in the atorvastatin group versus the placebo group at week 0; mean levels were similar at week 6. After the addition of atorvastatin, levels were significantly reduced to week 12 versus placebo (figure 2D and see supplementary table S1, available online only).

Consistent with changes in cholesterol moieties, apolipoprotein A-1 levels rose in both groups by week 6 and did not change after the addition of atorvastatin or placebo. Apolipoprotein B levels were significantly reduced by the addition of atorvastatin but not placebo by week 12 (figure 2E,F and see supplementary table S1, available online only). The apolipoprotein B/apolipoprotein A-1 ratio was unchanged from weeks 0 to 6 and was reduced by week 12 only with atorvastatin.

Effects on RA disease activity parameters

By week 6, ACR20 response rates were similar between groups; with continued tofacitinib treatment, rates increased further (expressed as change from week 0) at week 12 after 6 weeks of atorvastatin treatment, and decreased marginally with placebo. ACR50 and ACR70 response rates were similar at week 6 between groups. By week 12, ACR50 response rates increased further in the atorvastatin group versus placebo; ACR70 response rates improved in both groups by a similar magnitude. As specified in the protocol, only descriptive statistics were applied and interpretation of these observations must be cautious, because the study was limited by its size and the open-label use of tofacitinib. Improvements were observed in all DAS28-4(ESR) parameters at week 6, with further improvements at week 12. Health assessment questionnaire–disability improved in both groups to week 6 with further improvement with atorvastatin by week 12. There was a change from week 0 in the mean CRP level in both groups at week 6, with little further change at week 12. Data for RA disease parameters are presented in table 2.

Table 2

Measures of efficacy on RA disease activity at weeks 6 and 12


In the 6-week open-label period, 82 AE were reported by 52 patients. In the 6-week double-blind period, 36 AE were reported by 21 (42%) patients receiving atorvastatin and 36 AE by 19 (40.4%) patients receiving placebo. The most common treatment-emergent AE by system organ class are presented in table 3. Three patients experienced serious AE (weeks 0–6, n=2; 54-year-old woman, bacterial pneumonia; 25-year-old woman, worsening arthritis; weeks 6–12, n=1; 74-year-old woman, pneumonia, placebo/tofacitinib group). No deaths were reported. Changes in laboratory parameters throughout the study were similar for both groups, except for the number of moderate-to-severe cases of neutropenia (atorvastatin/tofacitinib, n=0; placebo/tofacitinib, n=5). Elevations in transaminases or creatine phosphokinase were uncommon (table 3). Fewer patients had hypertension in the placebo/tofacitinib arm than in the atorvastatin/tofacitinib arm at the end of the open-label and double-blind periods (table 3). Patient weight remained stable for the study duration.

Table 3

Summary of safety and laboratory data


Elevated cardiovascular risk and attendant mortality in RA are only partly explained by conventional risk factors; additional inflammation-related pathways are hypothesised to promote vascular pathology.22 One such pathway probably modifies lipid biochemistry. Cytokine and metabolic biochemical pathways are intimately linked by virtue of the evolutionary necessity to cross-regulate metabolic and immune function. Accordingly, in inflammatory states such as RA, serum lipids are generally lower compared with the normal population. Moreover, lipid particle composition (eg, HDL) may be altered such that increased proportions of serum amyloid A levels may render normally protective HDL particles atherogenic.23 Consequently, interpretation of normal vascular risk surrogates, such as those in Framingham and Reynolds risk scores, is complex in RA.24 The systemic release of pro-inflammatory cytokines, including tumour necrosis factor α and IL-6, probably mediate some of these downstream effects on lipid structure and function through currently ill-defined mechanisms,25 and inhibitors of these cytokines have been shown to elevate lipid parameters.26 ,27 Therefore, a priori, inhibition of the JAK signalling crucial for the downstream effects of many pro-inflammatory cytokines implicated in RA is likely to modulate serum lipid levels in these patients.

Tofacitinib has demonstrated efficacy in phase 3 clinical trials for the treatment of RA accompanied by elevations of lipid levels, consistent with the foregoing prediction. This study sought to characterise lipid changes and evaluate the impact of atorvastatin administered in combination with tofacitinib. Tofacitinib induced rapid elevation of mean total, LDL and HDL-cholesterol, triglycerides and apolipoprotein A-1 concentrations, which were sustained in placebo recipients. In contrast, the addition of atorvastatin significantly reduced elevated LDL and total cholesterol, triglycerides and apolipoprotein B to sub-week 0 levels.

Notably, atorvastatin reduced mean LDL-cholesterol below tofacitinib pretreatment means and into the optimal target range (ATPIII: <100 mg/dl; European Society of Cardiology: <2.5 mmol/l). Total cholesterol, apolipoprotein B and triglycerides responded similarly; these changes were also associated with reduced LDL particle number. HDL-cholesterol and apolipoprotein A-1 levels increased with tofacitinib and continued to increase regardless of treatment group to study end. These data are reassuring in confirming that observed lipid elevations in the context of tofacitinib administration are readily reduced by standard lipid-lowering therapy, should lipid levels reach those warranting treatment.

It remains unclear whether changes (generally elevations) in lipid levels associated with immunomodulatory therapy are necessarily associated with increased vascular risk. The net effect of such interventions may comprise a balance of the beneficial effects of inflammation reduction, including the potential impact on lipid particle assembly and composition, with the risks attendant on elevated total and LDL-cholesterol and triglycerides. It is noteworthy that total cholesterol/HDL-cholesterol, LDL-cholesterol/HDL-cholesterol and apolipoprotein B/apolipoprotein A-1 ratios did not change significantly in the first 6 weeks of tofacitinib treatment. Intriguingly, the elevated LDL-cholesterol observed in tofacitinib recipients was associated with a reduction in the ratio of small/total LDL particles, regardless of concomitant atorvastatin. Although the clinical relevance of changes in LDL particle size is uncertain, it has been suggested that reductions in small LDL particles are associated with a lower risk of ischaemic events.28 Further studies to obtain a better understanding of the physiological effects of tofacitinib on lipid mediators in RA are warranted.

Tofacitinib therapy led to significant, clinically meaningful and rapid improvements in signs, symptoms and reductions in RA disease activity. The magnitude of such changes was similar to that in the phase 3 programme,17 ,18 providing a reassuring reference point for interpretation of this open-label study. Over the 6 weeks following the addition of placebo, signs and symptoms improved further in terms of ACR70 response rates. There appeared to be a numerical increase in ACR rates with the addition of atorvastatin, but this requires further investigation with higher patient numbers before drawing any firm conclusions. Statins exert potentially anti-inflammatory effects via modulation of several intracellular immune signal pathways and degradation of the lipid rafts essential for lymphocyte activation;29 a role for the suppression of synovitis ex vivo and in rodent arthritis models has been proposed.30 Moreover, a number of studies have reported clinical benefit with statins in RA and other autoimmune diseases.31–33 However, one study reported statin-associated accelerated arthritis development in a mouse model, potentially linked to immunomodulatory properties of statins.34 The effects of statins on RA development have not been evaluated in prospective clinical trials. Further mechanistic studies are necessary to explore formally potential interactions between statin signal targets and the JAK pathways inhibited by tofacitinib.

Overall, the safety profile of tofacitinib 10 mg was similar to that in previous phase 210–13 and phase 314–18 studies of tofacitinib in patients with active RA. The most frequent AE were infections and infestations, during treatment with tofacitinib alone and with the addition of atorvastatin or placebo, and the incidences were similar among groups during the double-blind treatment period. Larger patient samples are required to characterise fully the safety profile for the concomitant administration of atorvastatin and tofacitinib. The incidence of serious AE, including serious infections, was low. Expected changes in laboratory parameters were observed for tofacitinib 10 mg bid, including decreases in neutrophil counts.

In summary, the addition of atorvastatin at week 6 of 12 weeks’ treatment with tofacitinib significantly reduced the elevations observed in a variety of lipid moieties in patients with RA and did not reduce the efficacy of tofacitinib on RA disease activity.


The authors would like to thank the patients who were involved in this study, and the A3921109 investigators and study team. The authors would also like to thank Stephen Reilly (clinical project manager), and Allison Brailey (lead programmer). Editorial support was provided by Gary Dever, PhD at Complete Medical Communications and was funded by Pfizer Inc.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Handling editor Tore K Kvien

  • *For details of the study investigators, please see end of paper.

  • Contributors CAC, ZL, MJB, SHZ, JDB: study design; HYK, SHL, DM, YWS: data acquisition; ZL: statistical analysis; IBM, HYK, SHL, DM, YWS, CAC, ZL, MJB, AZ, SHZ, JDB: data interpretation and manuscript preparation. All authors were involved in the drafting and reviewing of the manuscript and approved the final version. IBM had final responsibility for the decision to submit for publication.

  • Funding Study design, data collection, analysis and interpretation of results were funded by Pfizer Inc. Pfizer employees participated in study design, data analysis and interpretation.

  • Competing interests CAC, ZL, MJB, AZ, SHZ and JDB are employees of Pfizer Inc. IBM has received consultancy fees or honoraria, and research grants, from Pfizer Inc. DM has received consultancy fees or honoraria, research grants and fees for participating in reviewing activities from Pfizer Inc. HYK, SHL and YWS report no conflicts of interest.

  • Ethics approval The final protocol, amendments and informed consent documentation were approved by the institutional review boards and independent ethics committees of the investigational centres.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Study investigators Korea: Dr Yeong-Wook Song; Dr Soo-Kon Lee; Dr Jae-Bum Jun; Dr Ho-Youn Kim and Dr Sang-Heon Lee. USA: Dr Steven D Mathews; Dr Sanford M Wolfe; Dr Joseph S Habros; Dr David A McLain; Dr Antony C Hou; Dr John P Lavery; Dr William J Shergy; Dr David R Mandel; Dr Michael J Fairfax and Dr Geneva L Hill.