Trajectory of lipid profiles in rheumatoid arthritis patients receiving tofacitinib: Evidence from the prospective CENTRA cohort

Abstract

Aims

The cardiovascular risk of tofacitinib has been highly concerned. In this study, we investigated the trajectory of lipid profiles in patients with rheumatoid arthritis (RA) after receiving tofacitinib.

Methods

Patients were recruited from the prospective CENTRA cohort of RA patients. The data of RA disease activity, triglyceride (TG), total cholesterol (TCHO), high‐density lipoprotein cholesterol (HDL), low‐density lipoprotein cholesterol (LDL) and ratios of TCHO/HDL and LDL/HDL were collected at baseline, week 4, 12, 24 and 52. Lipid and lipoprotein concentrations were compared between the tofacitinib exposure group and the non‐exposure group by propensity score matching (PSM).

Results

Totally 374 patients were enrolled, with 137 received tofacitinib and 237 not. After 4 weeks of tofacitinib treatment, the levels of TG, TCHO, HDL and LDL were increased. TCHO and LDL concentrations returned to baseline levels by weeks 24 and 52, respectively, while TG and HDL levels remained elevated throughout 52 weeks. After 1:1 PSM, 133 patients in tofacitinib exposure group and 133 in non‐exposure group were identified. Compared to the non‐exposure group, the tofacitinib‐exposure group showed significantly increased TCHO levels at week 4 and continuously higher TCHO levels till week 24; HDL and LDL levels were also elevated from week 12 to week 24.

Conclusions

Four‐week exposure to tofacitinib induced elevation of TG, TCHO, HDL and LDL in serum. LDL and TCHO returned to baseline levels at week 24 and week 52, respectively; nevertheless, TG and HDL displayed continued elevated levels. Tofacitinib exposure was generally associated with a more remarkable elevation of lipid profiles.

What is already known about this subject

• RA patients have a significantly elevated risk of cardiovascular disease (CVD).

• JAK inhibitors like tofacitinib effectively treat RA but are known to increase serum lipid levels shortly after treatment initiation.

• However, it was unclear if these elevated lipid levels would reverse spontaneously over time, creating uncertainty about the need for lipid‐lowering interventions.

What this study adds

• This study reveals that elevated TCHO and LDL levels from tofacitinib spontaneously revert to baseline by weeks 24 and 52, while TG and HDL remain elevated.

• Tofacitinib causes a more significant lipid elevation compared to csDMARDs.

• The key message is the importance of monitoring lipid profiles during tofacitinib treatment, though the clinical impact on CVD risk requires further investigation.

1. INTRODUCTION

Rheumatoid arthritis (RA) is a systemic autoimmune disease, and the risk of cardiovascular disease (CVD) has long been a great concern. Previous studies reported remarkable elevation in the prevalence of ischemic heart disease and heart failure, and approximately 50% of mortality attributed to CVD among the RA population. ¹ , ² , ³ In the pathogenesis of CVD, the elevation of cholesterol levels, particularly low‐density lipoprotein (LDL), has been recognized as a critical risk factor. ⁴

The management of RA predominantly depends on the utilization of disease‐modifying antirheumatic drugs (DMARDs), including conventional synthetic DMARDs (csDMARDs), biological DMARDs (bDMARDs) and targeted synthetic DMARDs (tsDMARDs). Earlier investigations have indicated the varying influences of distinct DMARDs on lipid profiles among individuals with RA. ⁵ , ⁶ , ⁷

Janus Kinase (JAK) inhibitors (JAKi) represent a category of tsDMARDs. Through inhibiting JAK activity and subsequent downstream signalling pathways, JAKi effectively mitigates the action of various inflammatory cytokines implicated in RA pathogenesis, thereby regulating inflammation and inhibiting structural changes of bone. ⁸ The efficacy of JAKi in RA has been proven for over a decade. In the 2019 European League Against Rheumatism (EULAR) recommendations for the management of RA, JAKi was advised for individuals failing to achieve the target with the first csDMARDs strategy and exhibiting poor prognostic factors, similar to bDMARDs. ⁹ But the safety of JAKi, particularly cardiovascular concerns, has garnered increasing attention recently. In the 2022 updated EULAR recommendations, JAKi may be considered alongside bDMARDs after insufficient response to csDMARDs, with emphasis on the assessment of cardiovascular risk factors. ¹⁰

Evidence from real‐world studies and meta‐analyses indicated that therapy involving JAKi did not confer an elevated risk of major adverse cardiovascular events (MACE). ¹¹ , ¹² Meanwhile, clinical trials revealed that RA patients treated with tofacitinib experienced elevated levels of both high‐density lipoprotein (HDL) and LDL. ¹³ , ¹⁴ Pooled data of 6 Phase 3 studies and 2 open‐label long‐term extension studies demonstrated the rise of serum total cholesterol (TCHO), HDL and LDL concentrations during the initial 1–3 months of tofacitinib therapy, exhibiting more substantial increases compared to the patients receiving placebo, methotrexate (MTX) or tumour necrosis factor (TNF) inhibitors and stabilized thereafter. ¹⁵ Nonetheless, the LDL/HDL ratio remained relatively stable throughout the treatment period.

Tofacitinib, a JAK1/JAK3 inhibitor, is the first tsDMARD available in the Chinese market. Whether elevated lipid levels can be spontaneously reverted to the states before tofacitinib treatment, and whether pharmacological intervention is needed, remains unclear. Therefore, we conducted this study to investigate the trajectory of serum lipid profiles among RA patients following tofacitinib administration and further analyse the predictors of lipid changes.

2. METHODS

2.1. Setting

2.1.1. Cohorts

We used data from the CENTRA (Collaboratively intENsive Treat‐to‐target in RA) cohort, a prospective cohort set up in October 2015. More details of the cohort have been described in our previous studies. ¹⁶ , ¹⁷ Data were collected from the inception of the cohort until January 2024. The RA patients who met the following inclusion criteria were divided into either the exposure group or the non‐exposure group based on whether they received tofacitinib treatment or not. In the exposure group, patients were treated with tofacitinib monotherapy or in combination with csDMARDs, while in the non‐exposure group, patients received csDMARDs only. The study was approved by the ethics committee of Peking University First Hospital, and informed consent was obtained from each patient. The PI for this paper is Zhuoli Zhang, and she had direct clinical responsibility for patients.

2.1.2. Participants

All the participants from the CENTRA cohort must meet the following inclusion and exclusion criteria, they were not qualified for this study.

1. Inclusion criteria

MTX was chosen as the comparator to isolate the effect of tofacitinib from other biologic agents. The exposure group included patients treated with tofacitinib as monotherapy or in combination with MTX, while patients who received MTX as their first and only treatment were allowed to be recruited into tofacitinib non‐exposure group.

The baseline was set at the initiation of tofacitinib treatment for the exposure group, and the time of the first csDMARDs administration for the non‐exposure group.

All patients in the exposure group remained on treatment throughout follow‐up. In the non‐exposure group, treatment was limited to csDMARDs only.

• 2. Exclusion criteria
  • 1)Lipid‐lowering medications were used during the follow‐up period.
  • 2)Pregnancies were reported or detected during the study period.

2.2. Data collection

We collected the following data at baseline and weeks 4, 12, 24 and 52.

2.2.1. Baseline demographics

Age, sex, RA disease duration, body mass index (BMI) and current and past smoking status were recorded.

2.2.2. Laboratory results and disease‐related clinical data at each visit

Erythrocyte sedimentation rate (ESR, mm/h), C‐reactive protein (CRP, mg/L), TG, TCHO, HDL and LDL are tested at each visit. Tender joint count (TJC) and swollen joint count (SJC) based on 28 joint count, patient's global assessment (PGA, 100 mm VAS), evaluator's global assessment (EGA, 100 mm VAS) and questionnaires concerning patient‐report outcomes are performed.

Disease activity scores based on 28 joint count and ESR (DAS28‐ESR), disease activity scores based on 28 joint count and CRP (DAS28‐CRP), the simplified disease activity index (SDAI) and clinical disease activity index (CDAI) are calculated to assess the disease activity. The formula used to calculate the composite disease activity scores with corresponding definitions of remission and LDA is as following ¹⁸ , ¹⁹ , ²⁰ :

• ○CDAI = TJC28 + SJC28 + PGA + EGA
• ○SDAI = TJC28 + SJC28 + PGA + EGA + CRP
• ○DAS28‐ESR = 0.56√[TJC28] + 0.28 √[SJC28] + 0.7× $\ln [\mathrm{ESR}$] + 0.014 × PGA
• ○DAS28‐CRP = 0.56√[TJC28] + 0.28√[SJC28] + 0.36× $\ln [\mathrm{CRP}+1$] + 0.014 × PGA + 0.96

Remission is defined as CDAI≤2.8 or SDAI≤3.3, or DAS28‐ESR or DAS28‐CRP < 2.6. LDA is defined as CDAI>2.8 and ≤10 or SDAI>3.3 and ≤11, or DAS28‐ESR or DAS28‐CRP ≥ 2.6 and ≤3.2.

2.2.3. Carotid ultrasound

The carotid ultrasounds are arranged at baseline, year 1 of treatment, and then at 2‐year intervals.

2.2.4. Medications

The usages of csDMARDs, bDMARDs, tsDMARDs and glucocorticoids (GCs) were recorded. The prescribed GCs were converted to prednisolone equivalent dose. The disease duration is defined as the time from the onset of joint symptoms to the time of enrolment.

2.3. Statistical analysis

The descriptive statistics contained means and standard deviation (SD), medians and interquartile ranges (IQR) or percentages for each relevant variable. The comparison of baseline characteristics and lipid/lipoprotein levels during follow‐up between the tofacitinib exposure and non‐exposure groups, and comparisons of lipid and lipoprotein levels before and after treatment in the same individuals were assessed by the t‐test or non‐parametric test. Comparison of categorical data was performed using the chi‐square test. Propensity score matching (PSM) was employed to address potential confounding factors, including age, gender, disease duration and SDAI at the initial visit, thus excluding the influence of confounders on lipid metabolism outcomes among different treatment groups. These variables were incorporated into the propensity score estimation model for 1:1 matching with a tolerance of 0.2.

All reported p‐values were two‐sided and only associations with a p < 0.05 were considered statistically significant. All statistical analyses were performed using SPSS 22.0. Graphs were created using GraphPad Prism 8.0.

3. RESULTS

3.1. Demographics and baseline clinical characteristics of patients

Among 479 RA patients in the CENTRA cohort, 374 patients were enrolled on the study. The patient enrolment process is shown in Figure 1.

FIGURE 1.

Flowchart of patient selection.

The baseline characteristics of 374 patients were summarized in Table S1. Their mean (SD) age was 51.8 (13.5) years, with 296 (78.9%) female patients. The median (IQR) disease duration was 36.5 (11.9, 102.9) months and the median (IQR) BMI was 22.6 (20.2, 25.1). There were 61 (16.3%) current smokers or ex‐smokers.

The median (IQR) TJC28, SJC28, PGA and EGA at baseline were 2 (1, 6), 1 (0, 4), 30 (20, 60) and 30 (10, 50), respectively. Median (IQR) ESR and CRP levels were 23 (10, 42) mm/h and 6.4 (2.8, 17.1) mg/L. As for the disease activity assessment, the median (IQR) values of DAS28‐ESR, DAS28‐CRP, CDAI and SDAI were respectively 3.9 (2.5, 5.2), 3.4 (2.4, 4.6), 11 (5, 21.3) and 12 (5.4, 23.7), indicating predominantly moderate disease activity (MDA). Median (IQR) TG, TCHO, HDL levels were 1 (0.7, 1.4), 4.4 (3.9, 5.1), 1.3 (1.1, 1.5) mmol/L, and the average (SD) LDL, TCHO/HDL ratio, LDL/HDL ratio were 2.7 (0.8) mmol/L, 3.6 (1.0), 2.2 (0.9), respectively.

There were 137 (36.6%) patients treated with tofacitinib and 237 patients treated with csDMARDs only, respectively comprising the tofacitinib exposure group and tofacitinib non‐exposure group.

3.2. Impact of tofacitinib on serum lipid levels in RA patients

Levels of lipid and lipoprotein at 4‐, 12‐, 24‐ and 52‐weeks after tofacitinib initiation were assessed (Figure 2), and corresponding changes from baseline were also calculated (Table S2). A varying degree of increase in TG, TCHO, HDL and LDL levels starting from 4 weeks after initiation of tofacitinib treatment was found. TCHO and LDL concentrations reverted to baseline levels by weeks 24 and 52, respectively, while TG and HDL levels remained elevated up to week 52. Additionally, the TCHO/HDL and LDL/HDL ratios exhibited significant elevation compared to baseline at weeks 24 and 52.

FIGURE 2.

Trajectory of serum lipid profiles in 137 tofacitinib exposure patients for 52 weeks.

3.3. Comparisons of serum lipid profiles between tofacitinib‐exposure and non‐exposure groups after PSM

At baseline, no statistically significant differences were observed in gender, age, disease duration, BMI and lipid levels between the two groups of patients, but individuals who received tofacitinib treatment exhibited higher disease activity, indexed by DAS28‐ESR, DAS28‐CRP, CDAI and SDAI, than the tofacitinib non‐exposure group (Table 1).

TABLE 1.

Clinical characteristics and lipid profiles in the unmatched and PSM cohorts.

	Unmatched			Propensity score matched
	Tofacitinib exposure n = 137	Tofacitinib non‐exposure n = 237	p	Tofacitinib exposure n = 133	Tofacitinib non‐exposure n = 133	p
Basic characteristics
Women, n (%)	108 (78.8)	188 (79.3)	0.910	104 (78.2)	99 (74.4)	0.471
Age, mean (SD), y	53.2 (13.4)	51.1 (13.6)	0.147	52.8 (13.3)	53 (13.1)	0.889
Age of onset, mean (SD), y	47.4 (14)	44.6 (14.8)	0.074	46.9 (13.8)	47.4 (14.4)	0.774
Disease duration, median (IQR), m	36.5 (8.4, 101.8)	36.5 (12.2, 107.3)	0.399	36.5 (9, 102.6)	30.1 (6.1, 91.4)	0.554
BMI, median (IQR), kg/m2	22.3 (20, 24.7)	22.8 (20.3, 25.2)	0.449	22.3 (20, 24.9)	22.5 (20.2, 24.6)	0.818
Current or past smoking, n (%)	14 (15.9)	43 (19.9)	0.517	14 (16.9)	12 (14.5)	0.831
Baseline disease activity measures, median (IQR)
PGA (0–100)	40 (20, 70)	30 (20, 50)	0.015	40 (20, 60)	40 (20, 60)	0.831
EGA (0–100)	30 (20, 60)	20 (10, 40)	<0.001	30 (20, 50)	30 (10, 50)	0.095
TJC28	3 (1, 8.5)	2 (0, 5)	0.002	3 (1, 7)	3 (1, 8)	0.899
SJC28	2 (1, 5.5)	1 (0, 3)	<0.001	2 (1, 5)	1 (0, 5)	0.066
ESR, mm/h	27 (14.5, 45)	21 (9, 38.5)	0.008	27 (14, 42)	25 (12.5, 44)	0.502
CRP, mg/L	8.9 (3.5, 23)	5.8 (2.5, 13.3)	0.017	8.6 (3.5, 21.7)	6.2 (2.8, 17.2)	0.353
Baseline disease activity scores
DAS28‐ESR, mean (SD)	4.4 (1.7)	3.7 (1.7)	<0.001	4.3 (1.6)	4.2 (1.8)	0.491
DAS28‐CRP, mean (SD)	4.0 (1.6)	3.4 (1.5)	<0.001	3.9 (1.5)	3.7 (1.6)	0.547
CDAI, median (IQR)	15 (7.3, 25.5)	9 (4, 17.9)	<0.001	15 (7, 23)	12 (5, 25.5)	0.349
SDAI, median (IQR)	16.5 (8.5, 26.4)	10.1 (4.3, 21)	<0.001	16.2 (8.1, 25.5)	13.3 (5.7, 27.5)	0.353
Lipid and lipoprotein profiles
TG, median (IQR), mmol/L	1(0.7, 1.3)	1(0.8, 1.5)	0.225	1 (0.7, 1.3)	1 (0.7, 1.4)	0.551
TCHO, median (IQR), mmol/L	4.5(3.9, 5.1)	4.4(3.9, 5.1)	0.729	4.5 (3.9, 5.1)	4.3 (3.9, 5)	0.513
HDL, mean (SD), mmol/L	1.3 (0.3)	1.3 (0.4)	0.707	1.3 (0.3)	1.3 (0.4)	0.372
LDL, median (IQR), mmol/L	2.6(2.2, 3.2)	2.6(2.1, 3.1)	0.801	2.6 (2.2, 3.2)	2.6 (2.2, 3.1)	0.792
TCHO/HDL, median (IQR)	3.5(2.8, 4.2)	3.5(2.9, 4.3)	0.611	3.5 (2.8, 4.2)	3.6 (2.9, 4.3)	0.406
LDL/HDL, median (IQR)	2.1(1.5, 2.6)	2.1(1.5, 2.7)	0.758	2.1 (1.5, 2.6)	2.1 (1.6, 2.7)	0.509

BMI, body mass index; PGA, patient's global assessment; EGA, evaluator's global assessment; TJC28, tenderness joint count based on 28 joints; SJC, swollen joint count based on 28 joints; ESR, erythrocyte sedimentation rate; CRP, C‐reactive protein; DAS28, disease activity score based on 28 joints; CDAI, clinical disease activity index; SDAI, simplified disease activity index; TG, triglyceride; TCHO, total cholesterol; HDL, high density lipoprotein; LDL, low‐density lipoprotein.

Based on the propensity score for each patient, 133 individuals in each group who met the common support assumption were identified, thereby forming two cohorts with matched propensity scores and well‐balanced key characteristics such as gender, age, disease duration and disease activity (Table 1).

As depicted in Figure 3, the overall trajectory of lipid levels over the entire follow‐up period exhibited similar patterns in both groups, but significant differences were observed at specific time points. Notably, at week 4, TCHO levels in the tofacitinib exposure group surpassed those in the non‐exposure group and continued to ascend until week 24. HDL and LDL levels in the tofacitinib exposure group exceeded those in the non‐exposure group from week 12 through week 24.

FIGURE 3.

Comparisons of lipids and lipoproteins between tofacitinib exposure and non‐exposure groups after propensity score‐matched.

The changes in lipid levels at each visit from baseline were subsequently compared between the two groups (Table 2). Consistently, the changes in lipid profiles from baseline were notably elevated among patients in the tofacitinib exposure group compared to the non‐exposure group. Specifically, compared to the non‐exposure group, the tofacitinib‐exposure group exhibited significantly higher increments in TCHO at week 4 (0.12 vs. ‐0.10 umol/L, p = 0.003), week 12 (0.36 vs. ‐0.04 umol/L, p < 0.001) and week 24 (0.26 vs. 0.07 umol/L, p = 0.005); significantly more increases of HDL at week 12 (0.17 vs. 0.04 umol/L, p = 0.001) and week 24 (0.14 vs. 0.08 umol/L, p = 0.005); higher elevations in LDL levels at week 4 (0.19 vs. 0.00 umol/L, p = 0.026) and week 12 (0.15 vs. ‐0.04 umol/L, p = 0.005). Nevertheless, no significant differences were observed in the elevated levels of any lipid relative to baseline at week 52 between the two groups. Elevations in TG levels, TCHO/HDL ratios and LDL/HDL ratios did not exhibit significant disparities between the two groups throughout the entire follow‐up period.

TABLE 2.

Comparisons of changes of lipids and lipoproteins at week 4, 12, 24 and 52 from baseline between tofacitinib exposure and non‐exposure groups.

	Week 4 (n = 85)			Week 12 (n = 183)			Week 24 (n = 192)			Week 52 (n = 147)
	Tofacitinib exposure n = 53	Tofacitinib non‐exposure n = 32	p	Tofacitinib exposure n = 87	Tofacitinib non‐exposure n = 96	p	Tofacitinib exposure n = 96	Tofacitinib non‐exposure n = 96	p	Tofacitinib exposure n = 65	Tofacitinib non‐exposure n = 82	p
ΔTG, median (IQR), mmol/L	0.12 (−0.1, 0.4)	0.1 (−0.2, 0.4)	0.693	0.09 (−0.1, 0.3)	0.09 (−0.1, 0.3)	0.862	0.04 (−0.1, 0.3)	−0.02 (−0.2, 0.3)	0.233	0.08 (−0.1, 0.4)	0.01 (−0.2, 0.2)	0.086
ΔTCHO, median (IQR), mmol/L	0.45 (0.2, 1.1)	0.21 (−0.3, 0.6)	0.003	0.36 (0, 0.7)	−0.04 (−0.3, 0.4)	<0.001	0.26 (−0.2, 0.7)	0.07 (−0.3, 0.4)	0.005	0.14 (−0.5, 0.7)	0.07 (−0.4, 0.5)	0.326
ΔHDL, mean (SD), mmol/L	0.18 (0.3)	0.10 (0.2)	0.158	0.17 (0.3)	0.04 (0.2)	0.001	0.14 (0.3)	0.08 (0.2)	0.005	0.12 (0.3)	0.05 (0.3)	0.150
ΔLDL, median (IQR), mmol/L	0.19 (−0.1, 0.6)	0 (−0.3, 0.4)	0.026	0.15 (−0.2, 0.5)	−0.04 (−0.3, 0.3)	0.006	0.09 (−0.3, 0.4)	−0.09 (−0.4, 0.3)	0.120	0.09 (−0.3, 0.5)	−0.04 (−0.4, 0.3)	0.219
TCHO/HDL, median (IQR)	3.43 (2.8, 4.3)	3.61 (3.1, 4.3)	0.520	3.43 (2.9, 4)	3.4 (2.9, 4.2)	0.842	3.37 (2.8, 3.9)	3.27 (2.7, 4.1)	0.806	3.24 (2.7, 4.1)	3.39 (2.7, 4.2)	0.497
LDL/HDL, median (IQR)	2.1 (1.5, 2.6)	2.18 (1.6, 2.7)	0.631	2.05 (1.6, 2.6)	2.02 (1.6, 2.5)	0.725	1.99 (1.5, 2.3)	1.9 (1.4, 2.4)	0.867	1.88 (1.3, 2.5)	1.98 (1.4, 2.5)	0.534

During 52 weeks of follow‐up, there was no significant difference in disease activity between patients of the two groups. Taking the SDAI as an example, at baseline, weeks 4, 12, 24 and 52, the median (IQR) SDAI were 16.2 (8.1, 25.5), 9.1 (3.5, 16.4), 4.1 (2.1, 9.2), 4.1 (1.8, 8.2) and 3.1 (1.1, 6.4) in the tofacitinib‐exposure group, while 13.3 (5.7, 27.5), 7.9 (2.3, 16.2), 4.2 (2.2, 10.3), 3.3 (0.9, 6.7) and 2.1 (0.3, 5.0) in the non‐exposure group (p > 0.05 for all).

3.4. CVD events and newly detected carotid atherosclerotic plaque during follow‐up

Throughout 52 weeks of follow‐up, there was only one patient, in the tofacitinib‐exposure group, experienced acute ST elevation myocardial infarction (anterior wall). This 70‐year‐old man subsequently underwent stent implant and balloon dilation in the middle branch of the anterior descending artery. He had a 19‐year history of RA, with conventional CVD risk factors, including hypertension and smoking. Before the attack of acute myocardial infarction, he had received tofacitinib for 24 weeks. There was no significant elevation in lipid levels compared to baseline without using lipid‐lowering drugs. No CVD events were observed in tofacitinib non‐exposure group. Carotid ultrasound was performed in all patients at baseline and week 52. Newly developed carotid atherosclerotic plaque was found in 16 patients at week 52, 6 in tofacitinib‐exposure group and 10 in non‐exposure group.

4. DISCUSSION

In this study, we showed the trajectory of lipid profiles in RA patients receiving tofacitinib in a real‐world cohort, with comparison to those non‐exposed to tofacitinib. We found varied degrees of elevation in TG, TCHO, HDL and LDL after the administration of tofacitinib. Notably, there was a tendency to revert to baseline levels spontaneously in TCHO and LDL, whereas TG and HDL exhibited sustained elevation. Generally, patients treated with tofacitinib exhibited a more significant elevation in lipid levels compared to those receiving csDMARDs only.

CVD is a prevalent comorbidity in RA patients, often representing a principal cause of mortality. ²¹ , ²² EULAR recommendations for managing CVD risk in RA patients emphasized the heightened CVD vulnerability among RA patients and advocated for controlling disease activity in order to lower CVD risk. ²³ Previous studies revealed the favourable impact of csDMARDs and bDMARDs in reducing CVD events among RA patients. ⁵ , ²⁴ , ²⁵ Elevated serum levels of LDL and diminished HDL are recognized as risk factors for CVD in the general population, and statins are proved to lower LDL levels and subsequently reduce CVD risk. ²⁶ , ²⁷ Investigations into the impact of DMARDs on lipid profiles of RA patients have been undertaken. The COBRA study revealed elevated cholesterol levels in individuals treated with SASP monotherapy or in combination with other csDMARDs. ²⁸ Tocilizumab, an interlukin‐6 receptor antagonist, has been associated with increases in TCHO, HDL and LDL levels in numerous investigations. ²⁹ , ³⁰ Findings from a meta‐analysis indicated a consistent elevation in TCHO and HDL levels after treatment of TNF‐α inhibitors, with a transient increase of LDL level in short‐term, and a tendency towards reduction in ratios of TCHO/HDL and LDL/HDL in short‐ and mid‐term. ³¹ These suggest that different DMARDs may have different effects on lipid profiles.

Our findings are consistent with previous trials showing that early lipid increases after treatment with tofacitinib, but uniquely demonstrate spontaneous normalization of LDL and TCHO by 52 weeks. RA patients usually exhibit elevated levels of TCHO, HDL, LDL, apolipoprotein (Apo) A and Apo‐B, predominantly at 4–6 weeks following initiation of tofacitinib. ¹³ , ¹⁴ , ³² , ³³ Moreover, in patients with psoriatic arthritis and ulcerative colitis, the administration of tofacitinib was associated with increased levels of TG, TCHO, HDL and LDL; however, no significant alteration in LDL/HDL and TCHO/HDL ratios. ³⁴ , ³⁵ The prescribing information of tofacitinib indicates the potential of dose‐dependent elevation in cholesterol levels, with the most discernible effects typically manifesting by week 6, prompting lipid monitoring from 4 to 8 weeks after therapy initiation. ³⁶ In the CENTRA cohort, patients receiving tofacitinib similarly demonstrated elevated TG, TCHO, HDL and LDL levels after 4 weeks of treatment. Additionally, LDL and TCHO levels exhibited voluntary regression to baseline values by weeks 24 and 52, respectively, in the absence of lipid‐lowering drugs, while TG and HDL levels remained elevated. Notably, higher ratios of LDL/HDL and TCHO/HDL, respectively, at weeks 24 and 52 compared to baseline were noticed. The association of tofacitinib exposure with alteration in lipid profiles suggests the importance of closely monitoring lipid and lipoprotein levels during tofacitinib therapy in RA patients.

Previous research indicated that IL‐6 can decrease lipid levels by upregulating very‐low‐density lipoprotein receptor expression. ³⁷ Tocilizumab can regulate lipid metabolism by increasing cholesterol efflux capacity and decreasing the expression of LDL‐receptors in liver cells. ³⁸ , ³⁹ The mechanism of tofacitinib on lipid metabolism likely parallels to tocilizumab, given their similar intervention of IL‐6 pathway. ⁴⁰ Moreover, Charles‐Schoeman et al revealed tofacitinib could reverse the increased cholesterol ester catabolism in RA patients, thereby increasing cholesterol levels. ⁴¹

There is evidence illustrating the association of CVD with systemic inflammation as well as RA disease activity. ²³ , ⁴² Changes in lipid profiles have also been correlated with disease activity scores and inflammatory index. ⁴³ In this study, comparisons of RA disease activity between tofacitinib‐exposure and non‐exposure groups at each follow‐up visit yielded no significant disparities, but levels of lipid profiles revealed notable differences between the two groups. Tofacitinib exposure was associated with distinct elevations in lipid levels as early as week 4, particularly evidenced by TCHO compared to the non‐exposure group. Discrepancies in HDL and LDL levels emerged at week 12, however disappeared by week 52. Besides, the trajectory of lipid profiles showed a similar trend. These findings imply that the alterations of serum lipid levels in RA patients were induced by tofacitinib, not solely attributable to changes in disease activity.

The cardiovascular safety of tofacitinib has become a considerable debate, particularly concerning the occurrence of MACE and venous thromboembolism (VTE). It has been demonstrated that tofacitinib could attenuate atherosclerosis by inhibiting inflammation and protecting endothelial cells, thereby implying a potential reduction in cardiovascular risk. ⁴⁴ , ⁴⁵ Among the 7061 participants enrolled in the long‐term safety assessment of tofacitinib, 85 experienced MACE, yielding an incidence rate of 0.4 per 100 patient‐years (95% CI 0.3–0.5). ⁴⁶ Meta‐analytic findings indicated no heightened risk of MACE associated with JAKi therapy. ¹¹ Real‐world data also showed comparable MACE incidence rates between tofacitinib and bDMATDs (adjusted HR = 0.6, 95% CI 0.30–1.18). ¹² Oral Surveillance study, a safety end‐point trial enrolling patients aged at least 50 years with at least one additional cardiovascular risk factor, ended up with failing to establish noninferiority for tofacitinib compared to TNF inhibitor concerning MACE and malignancy risks. ⁴⁷ Post hoc analysis underscored a higher risk among older patients (>65 years) and smokers.

Among all patients with a mean age of 51.8 years in our study, there was 1 case that experienced acute myocardial infarction after 24 weeks of tofacitinib exposure. But the patients with newly developed carotid atherosclerosis were roughly the same in tofacitinib‐exposure and non‐exposure groups. Previous studies have advocated for the TCHO/HDL and LDL/HDL ratios as superior predictors of CVD compared to a single lipid marker. ⁴⁸ , ⁴⁹ Our investigation showed the elevations in the ratios following tofacitinib administration. The trajectory of changes in different lipid components after tofacitinib treatment highlights the complexity of JAKi on lipid metabolism regulation. Although the level of LDL returned to the baseline level, which indicates that this component has a limited impact on the risk of long‐term atherosclerosis, the continuous increase in TG requires clinical attention, especially in patients with other cardiovascular metabolic risk factors. The overall cardiovascular impact of these opposite changes is still uncertain and requires further research. It would be best to combine the assessment of lipoprotein function and longer‐term cardiovascular outcome data. Until then, for RA patients receiving tofacitinib treatment, it is recommended to closely monitor the full lipid profile rather than just LDL. Given that the ratio of LDL/HDL remains stable, and the ratio of TCHO/HDL has returned to the baseline level by the 52nd week, these changes do not necessarily require immediate low‐dose statin therapy for all patients, unless their baseline cardiovascular disease risk is high.

Several advantages should be noted in this study. Firstly, we outlined the trajectory of lipid profiles in RA patients based on a prospective cohort following 52 weeks of treatment, providing reasonable evidence for lipid monitoring during tofacitinib exposure. Secondly, we addressed the spontaneous reversal of elevated lipids after administrating tofacitinib, thereby supporting rational clinical use. Thirdly, by comparing the trajectory of lipid profiles between patients exposed to tofacitinib and csDMARDs only, we revealed the distinct impacts of various DMARDs on lipid levels, assisting the optimized patient management.

There were several limitations in our study. Firstly, we found instances of spontaneous reversal of lipid levels, yet sample size constrained further subgroup analyses to explore predictive factors for such reversals. Secondly, the effects of different doses of tofacitinib were not able to be analysed due to limited follow‐up visits, but the daily dose of tofacitinib was ranged 5–10 mg in all patients in this study. Thirdly, we focused on the trajectory of lipid profiles in RA patients after tofacitinib initiation, but we are not able to analyse cardiovascular events based on 52 weeks of follow‐up and low incidence rates. Fourthly, although prior cardiovascular disease is recognized as important factors influencing lipid profiles, it was not included in the propensity score matching model due to incomplete data. Future studies should aim to collect comprehensive cardiovascular risk data for more robust adjustment. Fifthly, excluding patients who initiated lipid‐lowering therapy may have introduced selection bias, potentially underestimating the magnitude or duration of lipid elevation. Future studies should conduct data screening at the time of initiating treatment. Lastly, lifestyle also influences lipid levels, which we failed to take into consideration in this study. Future studies with larger sample sizes and extended follow‐up duration are warranted.

5. CONCLUSIONS

This real‐world study investigated the alterations of lipid profiles in RA patients exposed to tofacitinib. We found that 4‐week tofacitinib treatment was associated with increases in TG, TCHO, HDL and LDL, but LDL and TCHO reverted to baseline levels by weeks 24 and 52, respectively. Moreover, tofacitinib was associated with a more significant elevation of serum lipids compared to csDMARDs only. This highlighted the importance of close monitoring of lipid profiles during tofacitinib treatment. Future research should focus on determining whether these lipid alterations contribute to increased CVD risk and exploring the necessity for lipid‐regulating interventions.

AUTHOR CONTRIBUTIONS

Zhuoli Zhang: conceptualization, methodology, writing—reviewing and editing, supervision. Hong Huang: methodology, data curation, writing—original draft preparation, software. Yan Geng, Yong Fan and Juan Zhao: visualization, investigation. All authors read and approved the final manuscript.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflicts of interests.

Supporting information

Table S1. Baseline demographics, clinical characteristics and lipid profiles of 374 patients.

Table S2. Changes of serum lipids and lipoproteins during 52 weeks in 137 patients after exposing to tofacitinib.

Huang H, Geng Y, Fan Y, Zhao J, Zhang Z. Trajectory of lipid profiles in rheumatoid arthritis patients receiving tofacitinib: Evidence from the prospective CENTRA cohort. Br J Clin Pharmacol. 2026;92(9):3049‐3059. doi: 10.1002/bcp.70598

DATA AVAILABILITY STATEMENT

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.