Study reference

Study characteristics

Patient characteristics

Intervention (I)

Comparison / control (C)

Follow-up

Outcome measures and effect size

Comments

EULAR

Sepriano, 2023

Type of study: SR without meta-analysis

Databases searched: MEDLINE, Embase, Web of Science and The Cochrane CENTRAL Register of Controlled Trials (Central)

Setting and Country: Worldwide. For details, see article and supplementary files.

Included studies:

Studies were heterogeneous, precluding data pooling, and results are presented descriptively.

-59 observational studies

- 30 RCTs

Observational studies

- 51 assessed only 1 outcome, 8 addressed ≥2 outcomes

- 27 studies evaluating the risk of infections; 23 included patients on bDMARD’s; of which also patients on JAKi; 3 only patients on csDMARD’s; and 1 patients either on tofacitinib or on csDMARD’s.

- 9 studies evaluated the risk of malignancies with bDMARD’s; and one of these also with tofacitinib.

- 13 studies assessed the risk of MACE; with 11 including patients on bDMARD’s; 6 of which included patients on JAKi; and 2 had patients only on csDMARD’s.

- Intestinal perforations and neuroinflammatory events were assessed in two studies, each in patients on bDMARD’s.

- All-cause mortality with bDMARD’s was assessed in seven studies.

- Seven studies addressed withdrawals due to adverse events with bDMARD’s and one with JAKi.

- One study assessed any serious adverse events, another any adverse event, both in patients with bDMARD’s,

- one evaluated pregnancy outcomes in patients on bDMARD’s.

RCTs

- 11 RCTs evaluated bDMARD’s and 19 RCTs evaluated tsDMARD’s.

- Most RCTs were not designed, and therefore, not powered, to evaluate

safety outcomes.

- The incidence of major adverse events was low and, mostly comparable between active treatment, placebo or active comparator.

- The exception was the ORAL-Surveillance study, a non-inferiority trial in which patients ≥50 years old who failed methotrexate and had ≥1 cardiovascular risk factor were randomised to tofacitinib 5 mg two times per day, tofacitinib 10 mg two times per day, or TNFi (adalimumab or etanercept). The trial was designed to test whether the upper limit of the 95% CI around the risk ratio of MACE or malignancies for tofacitinib (5 mg and 10 mg two times per day combined) compared with TNFi, was below 1.8 (the non-inferiority question).

- In addition, the ENTRACTE trial, which had a similar design and non-inferiority margin, compared the risk of MACE (primary endpoint) between tocilizumab and etanercept.

Source of funding and conflicts of interest:

Funding European Alliance of Associations for Rheumatology.

Competing interests: An extensive list of competing interests is reported in the article.

Inclusion criteria SR:

- No language restrictions

- Publications from 1 January 2019 to 14 January 2022, as an update of the previous SLR.

- The literature search addressed the safety of DMARD’s.

- Participants were adults (≥18 years old) with

a clinical diagnosis of RA.

- Studies including patients with other diagnoses were eligible only if results from patients with RA were presented separately.

- Observational studies, namely cohort studies/registries with >50 cases were the

main study type.

- RCTs with a primary safety outcome were included. In addition, RCTs and long-term

extensions (LTEs), selected in the accompanying SLR addressing efficacy,9 were also included to assess the safety of drugs without, or with limited real-world data available.

- Included safety outcomes are mentioned in the outcome measures and effect size column.

- For the risk of infection by SARS-CoV-2, only studies published after 31 May 2021 (the limit date of an SLR informing EULAR recommendations

focusing on the topic) were considered.

Exclusion criteria SR:

Studies on glucocorticoids were excluded, as they were dealt with in a separate SLR.

Important patient characteristics at baseline:

For details, see article and supplementary files.

Groups comparable at baseline?

For details, see article and supplementary files.

Describe intervention:

The intervention was any DMARD

(csDMARD, bDMARD—including biosimilars—or tsDMARD),

including all drugs (chloroquine, hydroxychloroquine, leflunomide,

methotrexate, sulfasalazine, abatacept, anakinra,

adalimumab, brodalumab, certolizumab pegol, etanercept, golimumab,

guselkumab, infliximab, ixekizumab, mavrilimumab,

ocrelizumab, ofatumumab, olokizumab, otilimab, rituximab,

sarilumab, sirukumab, tabalumab, tocilizumab, ustekinumab,

apremilast, baricitinib, decernotinib, evobrutinib, fenebrutinib,

filgotinib, fostamatinib, peficitinib, ruxolitinib, tofacitinib, upadacitinib),

formulations and duration.

Describe control:

Studies were only eligible

if they included a comparator group (either another DMARD,

combination therapy, or the general population). Studies on

glucocorticoids were excluded, as they were dealt with in a separate

SLR.

End-point of follow-up:

Follow-up time per included study:

For details, see article and supplementary files.

For how many participants were no complete outcome data available?

For details, see article and supplementary files.

The following safety outcomes were considered:

infections

(including serious infections, opportunistic infections such as tuberculosis (TB) and herpes zoster (HZ), risk of infection by SARS-CoV-2)

malignancies

major adverse cardiovascular events (MACEs)

venous

thromboembolism (VTE) including pulmonary embolism/deep venous thrombosis

Furthermore:

Mortality,

changes in lipid levels, elevations of creatine

phosphokinase, impairments in renal function,

elevations of

liver enzymes,

haematological abnormalities,

gastrointestinal side effects,

demyelinating disease,

induction of autoimmune

disease,

teratogenicity,

fertility and pregnancy outcomes.

For detailed results per outcome, see article and supplementary files.

Risk of bias (high, some concerns or low):

For details, see article and supplementary files.

Brief description of author’s conclusion:

The safety profile of bDMARD’s was further demonstrated. Whether the difference in incidence of malignancies, MACE and VTE between tofacitinib and

TNFi applies to other JAKi needs further evaluation.

WHAT THIS STUDY ADDS

⇒ The risk of malignancies and major adverse cardiovascular events (MACEs) is similar, or even decreased, with bDMARD’s compared with conventional synthetic (cs)DMARD’s.

⇒ Malignancies and MACE occurred more with

tofacitinib than with TNFi in patients who had certain cardiovascular risk factors, especially in patients older than 65 years of age.

⇒ Herpes zoster is more common with JAKi than

with csDMARD’s or bDMARD’s.

⇒ Lower intestinal perforations are rare but occur more often with tocilizumab than with other

bDMARD’s.

GRADE (per comparison and outcome measure) including reasons for down/upgrading:

Not specified.

Study reference

Study characteristics

Patient characteristics ²

Intervention (I)

Comparison / control (C) ³

Follow-up

Outcome measures and effect size ⁴

Comments

Fang, 2022

Various relevant outcomes

Type of study:

retrospective cohort study

Setting/Source and country:

NHI Research Database (NHIRD) and Taiwan Death Registry (TDR). Taiwan.

Funding and conflicts of interest:

Funding from Chang Gung Memorial Hospital (CMRPG3K0041,

CMRPG3K0042).

No competing interests.

Inclusion criteria:

- Newly diagnosed RA (ICD-9-

CM: 714.0 and ICD-10-CM: M05-06) patients identified from 1995 to 2017. Only RA patients with a catastrophic illness certificate, issued

after a formal review by an expert panel commissioned by the NHI Administration, were enrolled.

- Patients that received JAKi or TNFi 2015-2017

Exclusion criteria:

- Never received cDMARD’s

- Never received JAKi and TNFi

- Aged<18 or aged>80

- Received TNFi before 2015 or after 2017

N total at baseline: 3179

Intervention: 822

Control: 2357

Important prognostic factors²:

age ± SD:

I: 56.02 (12.11)

C: 55.81 (12.68)

Sex:

I: % M 20.45

C: % M 21.91

Groups comparable at baseline?

Yes

Describe intervention (treatment/procedure/test):

JAKi (tofacitinib)

Describe control (treatment/procedure/test):

TNFi (etanercept, adalimumab, or golimumab)

Length of follow-up:

Follow-up was performed

from the first use of JAKis or TNFis until the first occurrence of the individual study outcomes, switch to other agents, or December 31, 2018, whichever came first.

Relevant measures and effect size:

Coronary heart disease

I Incidence (100 PY) 0.48 (0.20-0.92)

C Incidence (100 PY) 0.45 (0.26-0.64)

HR (95% CI) 1.03 (0.45-2.36)

P value .9463

Stroke

I Incidence (100 PY) 0.33 (0.11-0.72)

C Incidence (100 PY) 0.46 (0.27-0.65)

HR (95% CI) 0.75 (0.29-1.94)

P value .5519

Overall venous

thromboembolism

I Incidence (100 PY) 0.32 (0.11-0.70)

C Incidence (100 PY) 0.48 (0.29-0.68)

HR (95% CI) 0.65 (0.25-1.70)

P value .3810

Deep vein thrombosis

I Incidence (100 PY) 0.26 (0.07-0.62)

C Incidence (100 PY) 0.44 (0.26-0.63)

HR (95% CI) 0.57 (0.20-1.64)

P value .3010

Tuberculosis

I Incidence (100 PY) 0.29 (0.09-0.66)

C Incidence (100 PY) 0.49 (0.29-0.68)

HR (95% CI) 0.60 (0.22-1.62)

P value .3122

Malignancy

I Incidence (100 PY) 0.39 (0.14-0.80)

C Incidence (100 PY) 0.35 (0.18-0.51)

HR (95% CI) 1.10 (0.44-2.78)

P value .8353

All-cause mortality

I Incidence (100 PY) 1.17 (0.64-1.70)

C Incidence (100 PY) 1.29 (0.97-1.61)

HR (95% CI) 0.91 (0.54-1.52)

P value .7099

Article conclusion:

Comparable safety issues and mortality rates were observed for JAKis and TNFis in RA patients treated in real-world settings.

Sensitivity analyses: The incidence rates of safety outcomes (CHD, stroke, OVT, DVT, TB, THR, TKR, malignancy, and all-cause

mortality) were higher in

those aged ≧50 than in the entire sample. After PSSW, an insignificant difference in these incidence rates between the 2 medication groups was seen in those aged ≧50.

Frisell, 2023

Various relevant outcomes

Type of study:

retrospective cohort study

Setting/Source and country:

ARTIS safety monitoring programme linking individual-level

longitudinal data on treatments, disease activity and other

clinical measurements from the Swedish Rheumatology Quality

Register (SRQ) to prospectively collected data in Swedish national healthcare registers ( National Patient Register, Prescribed Drug Register, census /taxation registers).

Sweden

Funding and conflicts of interest:

TF was supported by the Swedish Research Council (2016-00398,

2021-01418). JA was supported by the Swedish Research Council, Region Stockholm/Karolinska Institutet (ALF), the Swedish HeartLung-Foundation, the Swedish Cancer

Society, Nordforsk and Vinnova.

Inclusion criteria:

all patients with RA in Sweden who were recorded as starting any b/

tsDMARD between 1 January 2010 and 31 December 2020.

All approved b/tsDMARD’s used for RA in Sweden during the study period were included.

Exclusion criteria:

Drugs with fewer than 200 treatment episodes (here: anakinra (n=84) and upadacitinib (n=105)) were excluded from further analysis.

N total at baseline: 20 117 unique patients with RA (contributing a total of 34 279 treatment episodes)

Intervention:

JAKi:

BAR 1837

TOFA 426

bDMARD’s:

ADA 5526

IFX 2971

CZP 2179

GOL 1889

ABA 3434

RTX 4220

TCZ 2757

SAR 292

Control: ETA 8748

Important prognostic factors²:

age ± SD:

I: JAKi: BAR 61 (14); TOFA 59 (13)

bDMARD’s: 58 (14); 58 (14); 56 (15;) 57 (14); 61 (13); 64 (13;) 59 (14); 59 (14)

C: 58 (14)

Sex:

I: % M BAR 18; TOFA 18; bDMARD’s between 26 and 20%

C: % M 23

Groups comparable at baseline?

No, but weighting successfully balanced mean patient characteristics across the major treatment groups, it was not possible to reach acceptable balance for all variables in the two smallest groups (sarilumab, tofacitinib). Adjustment directly in multivariable Cox regressions gave very similar estimates throughout.

Describe intervention (treatment/procedure/test):

JAKi:

baricitinib, tofacitinib and upadacitinib.

TNFi: adalimumab, certolizumab pegol,

golimumab

and infliximab.

Other bDMARD’s:

abatacept, anakinra,

rituximab, sarilumab, tocilizumab

Describe control (treatment/procedure/test):

TNFi: etanercept,

Length of follow-up:

- Max.l study duration 1 January 2010 - 30 June 2021.

- Patients were considered exposed to a treatment from

their first ever start of that specific b/tsDMARD, as recorded in the SRQ, until treatment switch or discontinuation.

- When a patient switched or discontinued treatment, a lag time of 90 days was added after the treatment was stopped (183 days for rituximab) to capture adverse events linked to treatment discontinuation

but registered with some delay. If restarted within 90 days (183 days for rituximab), the two treatment episodes were merged.

-Swithching between biosimilars was not considered as discontinuation.

- Patients could contribute

with multiple treatment episodes on different drugs, but only the first ever start for each molecule.

- Follow-up was censored at death or first emigration from Sweden after treatment start.

Loss-to-follow-up and incomplete outcome data:

Data were complete on treatments, outcomes and most covariates derived from national registers, but about 30% lacked data on baseline DAS28 and HAQ. Missing covariate data were accounted for by multiple imputation.

Due to differences in market entry, the average FU pp was +/- 3 yrs for most bDMARD’s, below 2 yrs for JAKi & lowest for SAR, 1.3 yrs.

Relevant measures and effect size:       Only results for JAKi BAR and TOFA vs ETA are presented here. The article also includes results for other TNFi/bDMARD’s vs ETA.

Article also mentions results for other outcomes (i.e., treatment discontinuation due to adverse events, liver disease, depression, suicide, hospitalisation)

Crude IR/1000 PY; weighted IR/1000 PY; adjusted HRs (95%CI) vs etanercept

MACE

BAR 10.7; 9.9; 0.83 (0.49-1.42)

TOFA 12.9; 9.2; 0.78 (0.31-1.99)

ETA 10.1; 12.3; 1.0 (ref)

Serious infection

BAR 39.5; 33.6; 0.98 (0.75-1.29)

TOFA 47.3; 30.5; 0.89 (0.47-1.69)

ETA 24.8; 30.8; 1.0 (ref)

Diagnosed herpes zoster

BAR 10.0; 9.8; 3.82 (2.05-7.09)

TOFA 14.1; 10.2; 4.00 (1.59-10.06)

ETA2.0; 2.5; 1.0 (ref)

All-cause mortality

BAR 18.8; 22.0; 2.27 (1.51-3.40)

TOFA 8.7; 10.5; 1.09 90.40-2.92)

ETA 8.8; 12.1; 1.0 (ref)

Article conclusion:

- Data from ARTIS supports that the b/tsDMARD’s currently used to treat RA have acceptable and largely similar safety profiles, but differences exist in particular concerning tolerability and specific infection risks.

- few significant differences were observed for the serious adverse events under study. Neither CV events nor general serious infections were more frequent on BAR or TOFA versus bDMARD’s, but JAKi were associated with almost four times higher rates of hospital-treated HZ (vs ETA).

- Mortality rate was similar across TNFi, but about

30%–40% higher on the non-TNFi bDMARD’s, with the numerically highest HR seen for baricitinib.

NB: Low number of events limited some comparisons, in particular for sarilumab and tofacitinib.

Sensitivity analyses:

- Restricting the study period to the time after JAKi market entry reduced sample sizes drastically, and several differences between non-TNFi

bDMARD’s and etanercept were no longer significant,

but it did not materially alter the comparison between the

JAKi and etanercept.

- Ending follow-up at the onset of the COVID-19

pandemic had little impact on incidence rates or contrasts between bDMARD’s, but the reduced sample size for sarilumab and JAKi excluded them from most comparative analyses for these groups.

Hirose, 2022

Various relevant outcomes

Type of study:

multicentre, longitudinal observational study with both retrospective and prospective data

Setting and country:

12 hospitals and clinics for rheumatology in Japan

Funding and conflicts of interest:

No specific funding was received from any bodies in the public, commercial or not-for-profit sectors to carry out the work

described in this article.

Five out of the ten authors have an extensive list of conflicts of interest / liaisons with pharmaceutical companies.

Inclusion criteria:

- patients aged ≥20 years

-who fulfilled the 2010

ACR/EULAR classification criteria for RA

-who started treatment with TOF or ABT between January 2015 and January 2021.

- who had disease activity that was not controlled by MTX or csDMARD’s, or were unable to be treated with csDMARD’s, including MTX

- 2015-2017 retrospective data from patients’ medical records

- 2018-jan. 2021 prospective data

Exclusion criteria: non stated.

Only that prior use of bDMARD’s or JAKi was not an exclusion criterium.

N total at baseline: 370

Intervention: 187

Control: 183

Important prognostic factors²:

age ± SD:

I: 66.2 (11.5)

C: 70.7 (11.3)

Sex:

I: % M 15.5

C: % M 15.8

Groups comparable at baseline?

After IPTW, yes

Describe intervention (treatment/procedure/test):

Tofacitinib (TOFA)

Patients with an estimated glomerular filtration rate >60 ml/

min/1.73m2 received 5mg of TOF orally twice daily, whereas those with estimated glomerular filtration rate <60 ml/min/

1.73m2 received 5mg of TOF orally once daily.

Describe control (treatment/procedure/test):

Abatacept (ABT)

ABT was administered as an i.v. infusion (500 mg for patients weighing

<60 kg, 750mg for 60–100 kg, and 1000mg for >100 kg) at

weeks 0, 2 and 4, then every 4 weeks thereafter. Alternatively,

patients received 125mg by s.c. injection once weekly.

Length of follow-up:

52 weeks

Discontinuation reasons at week 4, 12, 26 and 52

Loss-to-follow-up:

Intervention:

N (%) 32 (17)

Reasons (describe): 2 consent withdrawal, 2 hospital transfer, 12 lack of efficacy, 1 concern for drug interaction, 1 clinical remission, 1 abdominal pain, 1 headache, 1 infectious arthritis of elbow, 1 non-tuberculous mycobacteria disease, 1 fever, 1 diverticulitis of the colon, 1 infectious enteritis, 3 herpes zoster, angina pectoris, 1 bacterial pneumonia, 1 herpes simplex, 1 generalized eczema, 1 elevation of liver enzymes.

Control:

N (%) 37 (20)

Reasons (describe):

5 consent withdrawal, 2 hospital transfer, 10 lack of efficacy, 3 clinical remission, 1 nausea, 1 skin rash, 1 cough, 1 fatigue, 2 gastrointestinal bleeding, 1 skin ulcer, 1 bacterial pneumonia, 1 COVID-19 pneumonia, 1 vertebral fracture, 1 bronchial asthma, 1 infectious enteritis, 1 pulmonary non-tuberculous mycobacterial infection, 1 eczema, 1 malignant lymphoma, 1 stomatitis, 1 gynecomastia

Incomplete outcome data:

The last observation carried forward method was used for patients who discontinued treatment before week 52 to include all patients in the analysis.

Relevant measures:

Serious infection, n (%)

I: 4 (2.1)

C: 4 (2.2)

P-value: 1.00

Herpes zoster, n (%)

I: 17 (9.1)

C: 5 (2.7)

P-value: 0.014

Cancer, n (%)

I: 0 (0)

C: 1 (0.5)

P-value: 1.00

MACE, n (%)

I: 2 (1.1)

C: 1 (0.5)

P-value: 0.49

VTE, n (%)

I: 0 (0)

C: 1 (0.5)

P-value: 0.49

Conclusion with regard to the relevant safety comparisons as stated in the article: “The incidence of herpes zoster was significantly higher in the TOF group than in the

ABT group (9.1 vs 2.7%, P=0.014). Three patients, including one case of grade 3, discontinued TOF prematurely owing to herpes zoster. No significant differences were noted in the incidence of serious infections, cancer or major adverse cardiovascular and venous thromboembolic events between the two treatment groups.”

Note: The primary

outcome was the remission rate at week 52 in each group. Analyses and results regarding the comparative effectiveness are described in more detail/length than safety.

Possible bias: between group differences in retrospective and prospective collected data? No information on numbers.

Also, in general, but especially for the prospective data, it is unclear how (on what grounds, taking into account data already collected retrospectively, matching on covariates?) patients were assigned to a treatment group.

Hoisnard, 2022

Cardiovascular outcomes (MACE, VTE)

Type of study:

nationwide population-based cohort study

Setting/Source and country:

French national health data system

France

Funding and conflicts of interest:

Funding statement: The authors have not declared a specific grant for this research from any

funding agency in the public, commercial or not-for-

profit

sectors.

Competing interests: None declared.

Inclusion criteria:

- All adults (≥ 18 years old) with at least one dispensation of a JAKi between 1 July 2017 and 31 May 2021 were eligible for inclusion in the exposed group.

- The unexposed population had to present RA severity comparable to that of the exposed cohort and use adalimumab.

- Patients with RA had recorded within the 5 years before the index date: (1) at least one hospitalisation with ICD-10 code as a principal, related or associated diagnosis specific for RA, and (2) long-term disease status with an ICD-10 code specific for RA.

- Patients who were JAKi-naive

and adalimumab-naïve (new users), defined as those who had not filled a prescription for one of these drugs for 1 year.

Exclusion criteria:

None specified in article.

N total at baseline:

Intervention: JAKi N=8481, Tofacitinib, N=3416,

Baricitinib, N=5065

Control: Adalimumab) N=7354

Important prognostic factors²:

age ± SD:

I: 59.3 (13.3)

C: 55.3 (13.4)

Sex:

I: % M 21.7

C: % M 28.8

Groups comparable at baseline?

No, but after IPTW yes.

Describe intervention (treatment/procedure/test):

JAKi, baricitinib 2 or 4 mg daily, tofacitinib 5 mg

or 10 mg two times per day

Describe control (treatment/procedure/test):

adalimumab

Length of follow-up:

Patients were followed up to the occurrence of each event

(MACE and VTE), death from any cause, exposure discontinuation or 31 December 2021, whichever came first.

Discontinuation date was the date corresponding to 120 days after the last reimbursement.

Only the first therapeutic sequence of a JAKi or

adalimumab was considered in this analysis.

Follow-up duration (days) Median (IQR)

JAKi 440 (203–846)

ADA 344 (185–686)

Loss-to-follow-up:

-

Incomplete outcome data:

-

Relevant measures and effect size:

MACE N events; IR (95%CI)/1000PY; N acute myocardial infarctions; N ischaemic strokes

JAKi 54; 4.3 (3.1 to 5.4); 29; 28

TOFA 14; 2.8 (1.4 to 4.3); 8; 6

BAR 40; 5.2 (3.6 to 6.8); 21; 22

ADA 35; 3.6 (2.4 to 4.9); 24; 11

Risk of MACEs JAKi vs ADA

HRw 1.0 (0.7 to 1.5) (p=0.99)

VTE N events; IR (95%CI)/1000PY; N pulmonary embolisms; N venous embolism and thrombosis

JAKi 75; 6.0 (4.6 to 7.3); 41 34

TOFA 29; 5.9 (3.7 to 8.0); 12 17

BAR 46; 6.0 (4.3 to 7.7); 29 17

ADA 32; 3.3 (2.2 to 4.5); 14 18

Risk of VTEs JAKi vs ADA:

HRw 1.1 (0.7 to 1.6) (p=0.63)

Article conclusion:

This study provides reassuring data regarding the risks of MACEs and VTEs in patients

initiating a JAKi versus adalimumab, including patients at high risk of cardiovascular diseases.

Subgroup analyses:

After weighting with the IPTW

method, risk of MACEs and VTEs was not significant whatever the subgroup (JAKi subgroups, age subgroups with CV risk factor, gender subgroups).

Sensitivity analyses were also not significant (death as competing risk analysis, a conventional multivariate Cox model with covariates and time-varying MTX [as in main analyses], and analysis with a broader definition of MACEs).

Huss, 2023

Outcome malignancies

Type of study:

prospective observational cohort study

Source and country:

The Swedish Rheumatology Quality register (SRQ), the National Patient Register, the Longitudinal Database for

Insurance and Labor Market Studies (LISA), the Prescribed Drug Register, the Population and Cause of death register and the National Cancer register.

Funding and conflicts of interest:

- This work was supported by funding from the Karolinska Institute Region Stockholm funds (ALF), the Swedish Research Council, the Swedish Cancer Society and the Swedish HeartLung Foundation. TF was supported by the Swedish Research

Council (2021-01418). Funders had no impact on the design or interpretation of the study or its results.

- Competing interests: VH and KH have no competing interests to declare. JA has

had or have research agreements with Abbvie, Astra-Zeneca, BMS, Eli Lilly, MSD, Pfizer, Roche, Samsung Bioepis, Sanofi, and UCB, mainly in the context of safety

monitoring of biologics via ARTIS/Swedish Biologics Register. TF and HB are partly employed by the ARTIS project.

Inclusion criteria:

-individuals above 18 years of age registered with RA or PsA in SRQ

- treatment initiations among patients with RA or PsA with: (1) TNFi, (2) non-TNFi

bDMARD or (3) JAKi

- included treatment initiations (TNFi and non-TNFi)

from 2016.

- For JAKi, treatment initiations from their market entry in 2017

Exclusion criteria:

- Because of the new-user

design, treatment episodes initiated before or ongoing at the start of the study period were not included (but contributed to the number of

previously used b/tsDMARD’s).

- participants who had a history of any previous cancer

other than NMSC were excluded from all analyses.

RA - N total at baseline: 10447

Intervention:

(1) 1967 JAKi

(2) 3520 non-TNFi

Control: (3) 7343 TNFi

PsA - N total at baseline: 4443

Intervention:

(1) 379

(2) 185

Control: (3) 4186

Important prognostic factors²:

RA

age (IQR):

I: (1) 59 (50–69)

(2) 60 (51–70)

C: (3) 56 (46–67)

(4) 57 (47–68)

Sex:

I: % M

(1) 18

(2) 21

C: % M

(3) 22

(4) 22

PsA

age (IQR):

I: (1) 52 (45–61)

(2) 54 (45–62)

C: (3) 50 (41–59)

(4) 49 (41–59)

Sex:

I: % M

(1) 29

(2) 29

C: % M

(3) 45

(4) 46

Groups comparable at baseline?

Not on all possible confounders, but analyses were corrected for all possible confounders.

Describe intervention (treatment/procedure/test):

(1) JAKi (baricitinib, tofacitinib, upadacitinib).

(2) non-TNFi

bDMARD (rituximab, abatacept, tocilizumab, sarilumab)

Describe control (treatment/procedure/test):

(3) TNFi (adalimumab, certolizumab pegol, etanercept, golimumab, infliximab)

(4) General population (For each patient, five general population

comparator subjects were identified, free from RA/PsA at

their index case’s first registration of RA/PsA and matched on

age, sex and region of domicile, from the Population register)

Length of follow-up:

- Start of follow-up was defined as the date of treatment start with each b/tsDMARD’s during the study period.

- Authors used an ever-treated

approach, in which each patient in each treatment cohort was followed from every treatment initiation until the occurrence of the outcome, death, emigration from Sweden or end of the study period on 31 December 2020.

- The total person time at risk in the JAKi, non-TNFi bDMARD and TNFi cohorts was 585, 418 and 12 623 patient years, respectively. The median follow-up times were 1.52, 2.25 and 2.44 years.

Loss-to-follow-up:

-

Incomplete outcome data:

-

Hazard Ratio analyses were adjusted for (1) age, sex, line of therapy; (2) comorbidities and SES => for these variables, there were no missing values (3) for disease-related factors, => missing values => complete case approach.

Imputation models were adjusted for all covariates included in the analysis model plus the event indicator and the Nelson-Aalen

estimate of the cumulative hazard.

Outcome measures and effect size:

- Main outcome: All cancers other than NMSC. And authors defined 19 sites of cancer as 19 individual outcomes.

- Authors only performed comparative analyses (i.e., estimated HRs) where the number events in each cohort were ≥5.

- In the JAKi cohort, there were 7 breast cancers (HR=0.73, 95% CI 0.29 to 1.86, vs TNFi), 6 haematopoietic (HR=1.90, 95% CI 0.70 to 5.16) and 7 lung cancers (HR=1.15, 95% CI 0.57 to 2.32).

- For all other sites other than NMSC, there were less than five events observed among the JAKi treated and thus are not presented.

Beneath, per outcome, per cohort: Events; PY, no (only for RA patients); Crude IR/1000 PY; Standardised IR/1000 PY; Fully adjusted HR (95%CI)

RA patients

Outcome:

All cancers other than NMSC

Cohort:

All JAKi

38; 3996; 9.5; 8.3; 0.94 (0.65 to 1.38)

Tofacitinib

8; 793; 10.1; 11.2; 1.08 (0.52 to 2.24)

Baricitinib

30; 3187; 9.4; 8.0 ; 0.92 (0.61 to 1.38)

Non-TNFi bDMARD

141; 11 051; 12.8; 10.5; 1.12 (0.88 to 1.43)

TNFi

213; 21 122; 10.1; 10.1; 1.0 (Reference)

General population

1245; 128 224; 9.7; 9.2; n/a

Outcome: NMSC

Cohort:

All JAKi

59; 3954; 14.9; 12.9; 1.39 (1.01 to 1.91)

Tofacitinib

11; 781; 14.1; 14.9; 1.56 (0.83 to 2.92)

Baricitinib

48; 3157; 15.2; 12.5; 1.37 (0.97 to 1.92)

Non-TNFi bDMARD

126; 11 027; 11.4; 9.0; 1.00 (0.78 to 1.28)

TNFi

189; 21 083; 9.0; 9.0; 1.0 (Reference)

General population

852; 128 630; 6.6; 6.2; n/a

PsA patients

All cancers other than NMSC

JAKi

5; 8.6; 7.3; 1.88 (0.68 to 5.16)

Non-TNFi

2; 4.8; –; –

TNFi

73; 5.8; 5.8; 1.0 (Reference)

Gen population

317; 5.9; 6.0; n/a

NMSC

JAKi

8; 13.9; 11.7; 2.05 (0.79 to 5.31)

Non-TNFi

2; 4.8; –; –

TNFi

73; 5.8; 5.8; 1.0 (Reference)

Gen population

209; 3.9; 3.9; n/a

Article conclusion:

among individuals with RA or PsA, we found no evidence of an increased short-term risk of all cancers other than NMSC for patients initiating JAKi compared with TNFi, but the risk of NMSC may be increased, at least in patients with RA.

Note: Patients that switched to a second drug of the same exposure category within the study period contributed twice to that cohort.

Study included additional analyses:

(1) with fitted models by time since treatment initiation (≤1, 1–2, ≥2 years) by inclusion of an interaction term and thereby relaxed the proportional hazards assumption.

(2) analyses by previous use of b/tsDMARD’s (0, 1–2, ≥3). (3) with latency period of 90 days, so that only cancers diagnosed 90 days or later after treatment start would contribute to analyses. (4) restricted the main analysis to a CV-enriched subset of the JAKi, non-TNFi and TNFi cohort

(5) a sensitivity analysis using an on-drug

approach for all cancers other than NMSC and NMSC, respectively, for

patients with RA.

(6) a sensitivity analysis restricting the follow-up

to Feb 2020 (for patients with RA) because the study period ended in December 2020 and the COVID pandemic could theoretically affect the

results.

With regard to these additional analyses: They did not appreciably vary from or alter the (fully adjusted HR) results of the main analysis

Jeong, 2022

Infections (serious, herpes zoster)

Type of study:

Setting/Source and country:

Korean Health Insurance Review & Assessment Service database.

South Korea

Funding and conflicts of interest:

This study was supported by the Korea Health Technology R&D Project through the Korea Health Industry Development Institute funded by the Ministry of Health Welfare, Republic of Korea (grant number HI14C1277).

The authors declare that they have no competing interests.

Inclusion criteria:

patients with seropositive RA who were prescribed bDMARD’s or tofacitinib between January 2010 and January 2019

Exclusion criteria:

- those who were prescribed bDMARD’s and tsDMARD in 2010 to limit the study population to new users

- patients with non-seropositive RA

- juvenile RA

- age <20 years

- HZ within 1 month of bDMARD or tsDMARD use

- HZ within 6 months of previous HZ

N total at baseline: 11720

Intervention: TOFA 701

Control: ABA 1153; ETA 2680

Important prognostic factors²:

age IQR:

I: TOFA 56 (47–63)

C: ABA 60 (52–68)

C: ETA 55 (45–64)

Sex:

I: % M TOFA 18.3

C: % M ABA 19.8

C: % M ETA 19.3

Groups comparable at baseline?

No, probably not but analyses were adjusted for potential confounders.

Describe intervention (treatment/procedure/test):

tofacitinib

infliximab,

adalimumab, golimumab,

tocilizumab, rituximab,

Describe control (treatment/procedure/test):

abatacept or etanercept

Length of follow-up:

Follow-up investigations

were performed for all patients until drug failure, development of HZ, or January 31, 2019, whichever

occurred earlier.

Patients with longer than 1 year of refill period from

the date of the last drug prescription were censored at the date of the last drug prescription. The drug failure-free survival was calculated for each individual to define

person-days of bDMARD’s or tsDMARD treatment for

the follow-up investigation.

Loss-to-follow-up:

-

Incomplete outcome data:

-

Patients with missing data, such as variables for adjustment, were excluded from the study.

Relevant measures and effect size:

Risk of HZ on RA patients during first bDMARD or tsDMARD use

IR/1000PY

TOFA 100.4

ABA 59.7

ETA 43.4

aHR (95% CI); p-value – ABA as reference

TOFA 2.46 (1.61–3.76); <0.001

ETA 1.19 (0.94–1.51); 0.146

aHR (95% CI); p-value - ETA as reference

TOFA 2.06 (1.38–3.08); <0.001

ABA 0.84 (0.66–1.06); 0.146

Risk of recurrent HZ on RA patients during first bDMARD or tsDMARD use

IR/1000PY

TOFA 185.6

ABA 83.8

ETA 70.7

aHR (95% CI); p-value – ABA as reference

TOFA 3.69 (1.77–7.69) <0.001

ETA 1.22 (0.75–2.00); 0.421

aHR (95% CI); p-value - ETA as reference

TOFA 3.01 (1.49–6.11); 0.002

ABA 0.82 (0.50–1.34); 0.421

After excluding patients who started bDMARD’s or tsDMARD before 2017 to normalize the clinical application period of TOFA in South Korea, TOFA (aHR, 3.26; 95% CI, 1.63–6.51; P < 0.001) showed an increased HZ risk compared to that of ABA.

Article conclusion:

Patients with seropositive RA treated with bDMARD’s or tsDMARD have a high HZ risk. In addition, HZ risk is significantly increased in RA patients with a history of HZ after the initiation of bDMARD’s or tsDMARD,

unlike that in the general population where prior HZ

may reduce the risk of recurrence. RA patients who

received tofacitinib showed higher overall, incident,

and recurrent HZ risks than those who received bDMARD’s.

Subgroup analyses:

subgroups of tofacitinib, infliximab, and adalimumab showed increased HZ risk compared to that of the corresponding abatacept subgroups. Higher HZ risk associated with Charlson comorbidity index > 2 (in ETA, IFX, ADA, TOFA) and with steroid use ≥ 5mg/day (in IFX, ADA, RTX, TOFA), for TOFA also with female gender, age above or below 65, more or less than 2 csDMARD’s.

Khosrow-Khavar, 2022

Cardiovascular outcomes (MACE, VTE)

Type of study:

observational cohort study

Setting/Source and country:

claims data from the Optum Clinformatics, IBM MarketScan, and Medicare databases.

USA

Funding and conflicts of interest:

This study was funded by internal sources of the Division of Pharma-coepidemiology and Pharmacoeconomics, Brigham and Women’s Hospital & Harvard Medical School, Boston, MA, USA.

Two authors received research grants for unrelated studies. All other authors have no conflict of interests to disclose.

Inclusion criteria:

- patients initiating treatment on tofacitinib or a TNFI (infliximab, adalimumab, certolizumab pegol, etanercept, and golimumab)

- Cohort entry date corresponded to first TNFI or tofacitinib dispensation (“index drug”) with a minimum of 365 days of continuous enrollment in health plan prior to and including the cohort entry date.

- Patients required at least two diagnosis codes for RA in any setting during the 365 days baseline period (between 7 and 365 days apart)

Exclusion criteria:

- TNFI users with a prescription of index TNFI and tofacitinib users with prescription of tofacitinib in the 365 days prior to cohort entry date;

- patients with a prescription of tofacitinib and TNFI on cohort entry date;

- patients missing data on age or gender, and those with admission to nursing facility or hospice on or prior to cohort entry date;

- TNFI users with history of use of any JAK inhibitor or with prescriptions for multiple agents from the TNFI class on cohort entry date;

- tofacitinib users with prescriptions of other approved JAK inhibitors (i.e. baricitinib or upadacitinib) on or at any point prior to cohort entry date.

From this source population of RA patients initiating treatment with tofacitinib or TNFI two cohorts were created:

(1) “real-world evidence (RWE)”, included all RA patients from routine care of ≥ 18 years of age in MarketScan and Optum (≥ 65 in Medicare) at cohort entry date.

(2) “RCT-duplicate cohort”, mimicked the inclusion and exclusion criteria of the ORAL surveillance trial. This was restricted to

- patients ≥ 50 years of age (65 in Medicare) with ≥ 1 methotrexate dispensation in six months prior to cohort entry date.

- and patients with at least one CV risk factor including history of smoking, hypertension, dyslipidemia, diabetes mellitus, ischemic heart disease, or family history of ischemic heart disease.

- Patients hospitalized with infections in the 30-days prior to cohort entry date and pregnant patients were excluded.

N total at baseline:

RWE cohort, 28568, 34083, and 39612 patients who met the inclusion and exclusion criteria were identified from Optum, MarketScan, and Medicare respectively of whom 13.2%, 15.6%, and 9.5% initiated treatment on tofacitinib

RCT-duplicate cohort, 6878, 8071, and 20121 patients were identified from Optum, MarketScan, and Medicare, respectively, of whom 11.6%, 14.3%, and 7.7% initiated treatment with tofacitinib.

Important prognostic factors²:

Data on covariates are provided per database per group for the RWE cohort.

Age; mean (std)

TOFA 56.8 (12.5)

TNFi 57.1 (13.2)

TOFA 54.7 (11.5)

TNFi 55.0 (12.0)

TOFA 72.1 (5.6)

TNFi 72.2 (5.6)

Female Gender; n (%)

TOFA 3043 (80.9)

TNFi 20046 (81.2)

TOFA 4333 (81.8)

TNFi 23503 (81.8)

TOFA 3134 (82.9)

TNFi 29819 (83.3)

Groups comparable at baseline?

Authors assessed 76 potential confounders and IPTW achieved excellent covariate balance in study populations with standardized differences close to zero for all covariates.

Describe intervention (treatment/procedure/test):

tofacitinib

Describe control (treatment/procedure/test):

TNFi (infliximab, adalimumab, certolizumab pegol, etanercept, and golimumab).

Length of follow-up:

Authors used an as-treated exposure definition whereby patients were followed from treatment initiation for study outcomes until treatment discontinuation or switch, insurance disenrollment, death, or end of the study period, whichever occurred first.

Total person years of follow-up:

RWE cohort

TOFA 11157

TNFi 80992

RCT-duplicate cohort

TOFA 3159

TNFi 29030

Loss-to-follow-up:

-

Incomplete outcome data:

-

Relevant measures and effect size:

Effect estimates were pooled across three databases using fixed effects model with inverse variance weighting.

Results for primary outcome: composite CV outcome consisting of hospitalizations for myocardial infarction or stroke.

RWE Cohort

Crude Incidence Rate (95% CI)/100PY

TOFA 1.31 (1.10 to 1.56)

TNFi 1.24 (1.16 to 1.33)

Crude Incidence Rate Difference (95% CI)/100PY (TNFi=ref):

0.20 (0.01 to 0.39)

Weighted HR (95% CI)(TNFi=ref):

1.01 (0.83 to 1.23)

RCT-Duplicate Cohort

Crude Incidence Rate (95% CI)/ 100PY

TOFA 1.83 (1.41 to 2.39)

TNFi 1.46 (1.32 to 1.61)

Crude Incidence Rate Difference (95% CI)/100PY (TNFi=ref):

0.46 (0.01 to 0.92)

Weighted HR (95% CI)(TNFi=ref):

1.24 (0.90 to 1.69)*

*which aligned closely with ORAL-surveillance results (HR: 1.33, 95% CI: 0.91 to 1.94).

Results for secondary outcomes in the RWE cohort: For individual CV outcomes, the pooled weighted HR (95% CI) was 1.04 (0.82 to 1.33) for MI, 0.93 (0.66 to 1.31) for stroke, 1.07 (0.79 to 1.46) for heart failure hospitalization, and 1.04 (0.78 to 1.40) for coronary revascularization (Supplemental Table 10) when comparing tofacitinib users with TNFI users. The pooled weighted HR (95% CI) was 1.20 (0.98 to 1.46) for all-cause mortality. For the positive control outcome, we successfully replicated the known association between tofacitinib and risk of herpes zoster (pooled weighted HR 1.98, 95% CI: 1.78 to 2.19).

Article conclusion:

Authors did not find evidence for an increased risk of cardiovascular outcomes with tofacitinib in RA patients treated in the real-world setting; however, tofacitinib was associated with an increased risk of cardiovascular outcomes, albeit statistically non-significant, in RA patients with cardiovascular risk factors.

Subgroup analyses in RWE cohort:

Some indication of possible increased risk of CV outcomes for CV history y/n and age subgroups: the pooled weighted HR (95% CI) was 1.27 (0.95 to 1.70) and 0.81 (0.61 to 1.07) among patients with and without history of cardiovascular disease respectively. The pooled weighted HR (95% CI) among patients ≤ 65 years of age was 1.00 (0.66 to 1.50) and 1.05 (0.84 to 1.33) for patient aged more than 65 years.

No TOFA-risk association was observed across other subgroups (gender, previous bDMARD). Consistent results were observed across other sensitivity and secondary analyses.

Sensitivity analysis in RCT-duplicate cohort:

In sensitivity analysis restricting comparator to adalimumab and etanercept (Supplemental Table 13), the pooled weighted HR (95%) CI for primary CV outcome was 1.32 (0.94 to 1.86).

Min, 2023

Various relevant outcomes

Type of study:

retrospective observational study

Source and country:

Korean National Health Insurance Service (NHIS) from the National Health Information Database (NHID) were used.

Funding and conflicts of interest:

This work was supported by the Konkuk University Medical Center Research Grant 2022 (grant No. K220112).

The authors have no conflicts of interest to declare for this study.

Set 1 newly diagnosed RA patients who were not exposed to either bDMARD’s or JAKis.

Set 2 included all RA patients, both bDMARD-naive and bDMARD-exposed.

Inclusion criteria set 1:

(1) patients who had an ICD-10 code for seropositive or seronegative RA as their first or second diagnosis between January 1, 2014, and December 31, 2020;

(2) used MTX for at least 6 months before initiation of JAKi or TNFi treatment.

Exclusion criteria set 1:

(1) RA patients were excluded who had an ICD-10 code for seropositive RA or seronegative RA between January 1, 2014, and December 31, 2016

(2) subjects were excluded who had a diagnosis code for AS, SLE, PS, PsA, Behçet’s disease, Crohn’s disease, or ulcerative colitis as their first or second diagnosis.

(3) patients with a main ICD-10 diagnosis code for AMI, stroke, VTE, ATE, or cancer 12 months prior to the initiation of TNFi or JAKi were excluded.

(4) Patients below 18 years

(5) RA patients who did not use JAKis or TNFis, or who started JAKis/TNFis before July 1, 2017, were excluded.

The inclusion and exclusion criteria for set 2 were identical to those of set 1; however, for set 2, the first exclusion step was not performed. In addition, in the last exclusion step for set 2, RA patients who started JAKis/TNFis before March 1, 2015, were excluded.

Set 1 & Set 2

N total at baseline:

Intervention: 645 & 2498

Control: 951 & 9267

Important prognostic factors²:

age ± SD:

I: 52.30±12.80 & 51.50±12.25

C: 50.16±14.44 & 52.00±13.27

Sex:

I: % M 24.5 & 16.8

C: % M 27.7 & 18.1

Groups comparable at baseline?

Not on all possible covariates, but analyses were corrected for all covariates.

Describe intervention (treatment/procedure/test):

JAKi (i.e., tofacitinib, baricitinib, or upadacitinib)

Describe control (treatment/procedure/test):

TNFi (i.e., etanercept, adalimumab, infliximab, or golimumab)

Length of follow-up:

All subjects were followed until December 31, 2021, or until each primary adverse event occurred, or if switching from JAKi to TNFi or TNFi to JAKi.

Loss-to-follow-up:

not described

Incomplete outcome data:

not described

The mean total follow-up duration was 2.1 years for the JAKi group and 2.5 years for the TNFi group in set 1, and 2.6 years for the JAKi group and 5.8 years for the TNFi group in set 2.

Outcome measures and effect size

- incidence rate ratio (95%CI): 1yr FU; total FU

- hazard ratio (95%CI), p-value: 1yr FU; total FU

Set 1

AMI, acute myocardial infarction

0.95 (0.41, 2.19); 0.77 (0.40, 1.48)

0.79 (0.33, 1.87), 0.59; 0.65 (0.34, 1.27), 0.21

Stroke

0.27 (0.06, 1.20); 0.44 (0.15, 1.33)

0.31 (0.07, 1.43), 0.13; 0.43 (0.14, 1.33), 0.14

cardiovascular related mortality

1.47 (0.43, 5.09); 1.43 (0.57, 3.63)

0.21 (0.04, 1.21), 0.08; 1.26 (0.47, 3.40), 0.64

MACE

0.71 (0.37, 1.34); 0.69 (0.42, 1.14)

0.64 (0.33, 1.22), 0.18; 0.59 (0.35, 0.99), 0.04

All-cause mortality

1.23 (0.38, 4.03); 1.46 (0.70, 3.03)

0.97 (0.29, 3.28), 0.97; 1.41 (0.65, 3.04), 0.39

VTE, venous thromboembolism

NA; 0.36 (0.08, 1.62)

NA; 0.33 (0.07, 1.54), 0.16

ATE, arterial thromboembolism

NA; 1.79 (0.11, 28.62)

NA; 1.58 (0.10, 25.47), 0.75

Cancer (excluding non-melanoma skin cancer)

0.80 (0.30, 2.17); 1.53 (0.81, 2.87)

0.68 (0.25, 1.88), 0.46; 1.29 (0.68, 2.45), 0.44

Non-melanoma skin cancer

NA; NA

NA; NA

Set 2

AMI, acute myocardial infarction

0.62 (0.41, 0.93); 0.69 (0.53, 0.90)

0.65 (0.43, 0.99), 0.04; 0.67 (0.51, 0.87), <0.01

Stroke

0.55 (0.31, 0.96); 0.74 (0.52, 1.05)

0.60 (0.34, 1.06), 0.08; 0.74 (0.52, 1.06), 0.10

cardiovascular related mortality

1.44 (0.92, 2.27); 1.14 (0.85, 1.54)

1.66 (1.05, 2.63), 0.03; 1.34 (0.98, 1.83), 0.07

MACE

0.74 (0.56, 0.98); 0.84 (0.70, 1.00)

0.81 (0.61, 1.06), 0.13; 0.80 (0.67, 0.97), 0.02

All-cause mortality

1.32 (0.86, 2.03); 1.19 (0.94, 1.52)

1.43 (0.93, 2.22), 0.10; 1.71 (1.32, 2.22), <0.01

VTE, venous thromboembolism

1.52 (0.88, 2.63); 1.35 (0.92, 1.97)

1.55 (0.90, 2.70), 0.12; 1.34 (0.90, 1.99), 0.15

ATE, arterial thromboembolism;

1.86 (0.34, 10.13); 0.87 (0.27, 2.86)

1.83 (0.33, 10.05), 0.49; 0.93 (0.27, 3.20), 0.91

Cancer (excluding non-melanoma skin cancer)

0.57 (0.38, 0.87); 1.02 (0.81, 1.29)

0.59 (0.39, 0.89), 0.01; 0.94 (0.74, 1.19); 0.61

Non-melanoma skin cancer

NA; 1.10 (0.33, 3.66)

NA; 1.38 (0.39, 4.88), 0.61

Article conclusion:

JAKis did not increase the risk of AMI, stroke, CV-related mortality, MACE, VTE, ATE, or cancer in Korean RA patients relative to TNFis.

In Set 1:

Only the total follow-up HR for MACE was significantly lower in the JAKi group compared to the TNFi group.

In Set 2:

1-yr follow-up HRs for AMI, Cancer (excl. NMSC), and total follow-up HRs for AMI and MACE were significantly lower in the JAKi group compared to the TNFi group while 1-yr follow-up HR for CV related mortality, and

the total follow-up HR for all cause mortality were significantly higher in the JAKi group compared to the TNFi group.

Mok, 2023

Various relevant outcomes

Type of study:

retrospective observational cohort study

Setting and country:

the Hong Kong Biologics Registry

Funding and conflicts of interest:

This study was funded by the Hong Kong Society of

Rheumatology, which has obtained a mini-grant from Pfizer.

Disclosure statement: C.C.M. received a one-off speaker’s honorarium from Pfizer at the APLAR 2022 meeting. H.S. received

speaker’s honorarium from Pfizer in the past 2 years.

Inclusion criteria:

patients who fulfilled the ACR/

EULAR criteria for RA and were ever treated with JAK and TNF inhibitors according to data from the Hong Kong Biologics registry between 2008 and December 2021.

Exclusion criteria:

none reported.

N total at baseline: 2471 courses in 1732 RA patients

Intervention: 551

Control: 1920

Important prognostic factors²:

age ± SD:

I: 57.9 (11.4)

C: 52.6 (12.6

Sex:

I: % M 18.1

C: % M 15.7

Groups comparable at baseline?

No, many significant differences, but analyses are corrected for covariates.

Describe intervention (treatment/procedure/test):

JAKi

Describe control (treatment/procedure/test):

TNFi

Length of follow-up:

Length of treatment/course of JAKi or TNFi (multiple courses per patient possible). Reasons for withdrawal, clinical inefficacy or SAE and mortality were outcome measures.

Loss-to-follow-up:

The cumulative withdrawal rates for JAKis and TNFis

at year 1, 3 and 5 due to clinical inefficacy were 22.4%, 36.9%, 43.5%; and 28.9%, 43.7%, 52.8%, respectively (log rank test; P¼0.001).

The corresponding withdrawal rates of the JAKis and TNFis due to SAEs were 8.6%, 14.5%, 18.5%; and 12.9%, 20.2%, 24.5%, respectively (log rank test;

P¼0.004).

The hazard ratios (HRs) of withdrawal of the JAKis relative to the TNFis for clinical inefficacy and SAEs

were 0.77 (95% CI: 0.63, 0.89) and 0.57 (95% CI: 0.42,

0.79), respectively, after adjustment for age, sex, duration of RA and the use of concomitant csDMARD’s.

Outcome measures and effect size:

Incidence/100PY JAKis; Incidence/100PY TNFi; Incidence ratio (95%CI); P

Cardiovascular events

0.63; 0.40; 1.67 (0.65, 4.32); 0.29

Cerebrovascular events

0.54; 0.36; 0.93 (0.36, 2.41); 0.89

Peripheral vascular events

0.18; 0.00; –; –

Total MACEs

1.34; 0.75; 1.49 (0.79, 2.84); 0.22

Venous thromboembolism

0.09; 0.02; 3.90 (0.20, 78.1); 0.37

Total cancer

0.81; 0.85; 0.65 (0.32, 1.35); 0.25

Total infections

16.3; 9.90; 1.24 (1.04, 1.48); 0.02

Total serious infections

8.20; 5.90; 0.97 (0.76, 1.23); 0.77

Death due to vascular events

0.09; 0.00; –; –

Death due to infections

0.18; 0.17; 0.70 (0.19, 2.62); 0.59

Death due to cancer

0.18; 0.15; 0.70 (0.15, 3.38); 0.66

All-cause death

0.72; 0.47; 0.78 (0.35, 1.74); 0.54

JAKi vs TNFi Univariate HR (95%CI), P

JAKi vs TNFi Multivariate HR (95%CI), P

MACEs

2.09 (1.10, 3.95), 0.02

1.36 (0.62, 2.96), 0.44

Cancers

0.84 (0.39, 1.83), 0.66

0.87 (0.39, 1.95), 0.74

Infections

1.43 (1.16, 1.77), 0.001

1.08 (0.84, 1.39), 0.55

Article conclusion:

In a real-life setting, there is no increase in MACEs or cancers in users of JAKis compared with TNFis. However, the incidence of

non-serious infections, including herpes zoster, was increased in users of JAKis.

No significant difference in:

- incidence of MACE after adjustment for age, sex and follow-up duration.

- incidence of new cancers. However, Users of the JAKis had numerically higher incidence of breast, upper gastrointestinal and haematological cancers whereas users of the TNFis had numerically higher incidence of lung, head and neck, genitourinary, and non-melanoma skin cancers compared with their counterparts.

- incidence of all-cause mortality.

- incidence of serious infections.

- incidence of TB, atypical mycobacterium, COVID-19

Significant difference in

- incidence of all infections => higher in the JAKi than in the TNFi group after adjustment for age, sex and RA duration, which was contributed by more genitourinary, lower respiratory tract, skin and soft tissue infections.

- HZ infection was significantly

more common with the JAKis than TNFis (3.49 vs 0.94 per 100 patient-years; P<0.001).

Molander, 2022

Cardiovascular outcomes (MACE, VTE)

Type of study:

nationwide register-based, active comparator, new user design cohort study

Setting/Source and country:

The Swedish Rheumatology

Quality Register was linked to national health registers;

Sweden

Funding and conflicts of interest:

This study has received funding from Swedish Research Council, the

Swedish Heart Lung Foundation, Nordforsk, Vinnova, and the Karolinska Institutet

Region Stockholm funds (ALF).

Karolinska Institutet, with JA as principal investigator, has

or has had research agreements with Abbvie, Astra-Zeneca,

BMS, Eli Lilly, Galapagos,

MSD, Pfizer, Roche, Samsung Bioepis, Sanofi, and UCB, mainly in the context of

safety monitoring of biologics via ARTIS/Swedish Biologics Register.

Inclusion criteria:

Not explicitly mentioned.

Authors included b/tsDMARD treatment initiations between 1 January 2010 and 31 December 2020 among all in the Swedish Rheumatology Quality Register (SRQ) registered patients with RA above 18 years of age.

Using validated algorithms applied to National Patient Register (NPR), authors further identified the entire RA population in Sweden (the ‘overall RA cohort’) defined by at least two separate visits listing RA at a rheumatology or internal medicine clinic before or during the

study period.

For each patient in b/ts cohorts five individuals from the general population were matched on age, sex and residential area.

Exclusion criteria:

Subjects with a VTE registered during the year prior to start of follow-up

were excluded.

N total at baseline: 27 610 unique patients with RA contributed with 32 737 b/tsDMARD’s treatment exposures

Intervention: 2150 JAKi

Control: 15090 TNFi

Important prognostic factors²:

age Median (IQR):

I: 60 (51–70)

C: 57 (47–67)

Sex:

I: % M 18

C: % M 23

Groups comparable at baseline?

No, but comparative analyses were adjusted.

Describe intervention (treatment/procedure/test):

JAKi (tofacitinib, baricitinib, upadacitinib)

Also:

rituximab, interleukin

6 inhibitors (IL6i) (tocilizumab, sarilumab), abatacept

Describe control (treatment/procedure/test):

TNFi (etanercept, infliximab,

adalimumab, golimumab, certolizumab pegol),

Length of follow-up:

Treatment initiation was defined as the registered date of treatment start in the SRQ. For each

of these treatment cohorts, we used a first initiation per molecule approach, meaning that one individual could contribute to each treatment cohort more than once, but only once with

each individual drug. Switch

from an originator to biosimilar was considered the same treatment

episode, as was restarting the same treatment within 90 days after discontinuation (180 days for rituximab) if no other bDMARD was initiated in between. An on-drug

approach was used. Treatment discontinuation was defined as either of

(1) registered date for discontinuation in the SRQ, (2) start of alternative b/tsDMARD in SRQ or (3) date of filled prescription for alternative b/tsDMARD from the PDR, whichever came first.

For the overall RA cohort, follow-up started at first day

of study period (or at the time point of the second ICD10 code registration for RA, if later). For the matched general population cohort, follow-up started at the time point of the first recorded

treatment initiation in the corresponding index individual with RA. For all cohorts, follow-up

ended at 60 days after discontinuation of the DMARD treatment in question, first VTE event,

death, emigration or end of study period (31 December 2021), whichever came first.

Loss-to-follow-up:

-

Incomplete outcome data:

-

For covariates with missing data they used the missing indicator method.

Relevant measures and effect size:

venous thromboembolism (VTE)

• Obs.; PYs at risk; VTE events; Standardised IR/1000 PYs (95% CI);
• unadjusted HR (95%HR); model 1 HR (95%HR); model 2 HR (95%HR); model 3 HR (95%HR)

JAKi vs. TNFi

• 2354; 4184; 48; 11.33 (8.54 to 15.04);
• 2.16 (1.59 to 2.93); 1.94 (1.40 to 2.70); 1.63 (1.17 to 2.28); 1.73 (1.24 to 2.42)

Baricitinib vs. TNFi

• 1825; 3412; 41; 11.35 (8.35 to 15.41);
• 2.27 (1.64 to 3.15); 2.00 (1.41 to 2.83); 1.69 (1.19 to 2.40); 1.79 (1.25 to 2.55)

Tofacitinib vs. TNFi

• 424; 667; 7; 11.30 (5.39 to 23.70);
• 1.96 (0.92 to 4.15); 1.91 (0.89 to 4.11); 1.56 (0.72 to 3.35); 1.66 (0.77 to 3.59)

TNFi

• 19950; 55765; 287; 5.15 (4.58 to 5.78)

Model 1 adjusted for age, sex and line of therapy.

Model 2 additionally adjusted for comorbidities and socioeconomic variables. Model 3 additionally adjusted for RA disease variables, civil status and smoking, using an indicator for missing variables.

pulmonary embolism (PE)

JAKi vs. TNF fully adjusted HR

3.21 (95% CI 2.11 to 4.88)

deep vein thrombosis (DVT)

JAKi vs. TNF fully adjusted HR

0.83 (95% CI 0.47 to 1.45)

Article conclusion:

Patients with RA treated with JAKi in clinical practice are at increased risk of VTE compared with those treated with bDMARD’s, an increase

numerically confined to PE.

Stratified analyses:

- analysed separately by sex, the incidence rates of VTE were approximately 50% higher in males than in females, and the HRs were numerically higher for males compared with females for those treated with IL6i versus TNFi as well as JAKi versus TNFi (2.66 (95% CI 1.44 to 4.92) for males and 1.53 (1.02–2.30) for females).

- stratified by history of VTE, the incidence rate of VTE was almost nine times higher

for individuals with a previous VTE versus those without, but all HRs for b/tsDMARD’s versus TNFi were close to 1. By contrast, among those without a previous VTE, the HRs were increased with IL6i versus TNFi and with JAKi versus TNFi (1.95 (1.33 to 2.87).

Sensitivity analyses

-altering the definition of VTE as well as the definition of follow-up showed similar results as the main analyses, although the HRs were generally lower when extending follow-up from 60 days after exposure-drug discontinuation until start of (any) next treatment

- limiting the study period to 2016–2021, HRs were very similar to those in the main analysis.

Additional analyses

- analysing the subgroup of those fulfilling an emulation of the inclusion and exclusion criteria of the ORAL surveillance study, the incidence rates of VTE were around 50% higher,

although the HRs for JAKi versus TNFi were similar or lower compared with the main analysis

Pina Vegas, 2022

Cardiovascular outcomes (MACE, VTE)

Type of study:

Observational cohort study

Setting/Source and country:

the French National

Health Insurance database linked with the national hospital discharge database.

France

Funding and conflicts of interest:

There are no funders to report for this submission.

Disclosure statement: L.P.V., P.L.C., L.P., M.P. and E.S. have no conflict of interest to declare. P.C. has received consulting fees from Abbvie, Pfizer, Roche-Chugai, Bristol-Myers Squibb, MSD, UCB, Novartis, Janssen, Lilly and Celgene (<$10 000 each), and has been an investigator for Roche Chugai, Sanofi Aventis, Celgene, Pfizer, MSD, Novartis and BMS.

Inclusion criteria:

- Adults (≥18 years old) with PsA,

ICD-10 code M07] registered in the SNDS between 2015 and 2019

- patients with at least one prescription of bDMARD or

apremilast for PsA

- bDMARD- and apremilast-naive patients, defined as those who had not filled a prescription for one of these drugs for 1 year.

Exclusion criteria:

patients with a history of acute myocardial infarction, unstable angina, chronic ischaemic heart disease, ischaemic stoke or transient ischaemic attack identified within 5 years before the index date

N total at baseline:

Intervention: 1885 apremilast

Control: 7289 TNFi (total bDMARD’s included, n=9510)

Important prognostic factors²:

age ± SD:

I: 54.0 (12.5)

C: 48.2 (12.8)

Sex:

I: % M 44.3

C: % M 41.2

Groups comparable at baseline?

No, but IPTW was applied.

Describe intervention (treatment/procedure/test):

tsDMARD: Apremilast

IL-12/23i

IL-17i: secukinumab and ixekizumab

Describe control (treatment/procedure/test):

TNFi: etanercept, infliximab, adalimumab, certolizumab and

golimumab

Length of follow-up:

Exposure to a molecule was defined as the time from initiation to discontinuation. We defined the discontinuation of treatment as (i) a period of >90 days without a dispensation of the same treatment after the period covered by the previous reimbursement [28], or (ii) a switch of systemic treatment (other bDMARD or apremilast). The period covered by a prescription was 30 days for all molecules except infliximab (56 days) and ustekinumab

(82 days). The discontinuation date was defined as the end of the 90-day period, and the switch date was defined as the date on which another systemic treatment was first reimbursed. Only the first therapeutic sequence of bDMARD or apremilast was considered in this analysis.

In the intention-to-treat analysis, patients were followed up to the MACE event, death from any-cause, systemic treatment switch, lost to follow-up (defined by the absence of any reimbursement for 12 consecutive months) or 31 December 2019, whichever came first.

Loss-to-follow-up:

-

Incomplete outcome data:

-

Relevant measures and effect size:

N(%) MACE, IR/1000PY(95%CI)

TNF inhibitors (n¼7289) 30 (0.4) 2.8 (1.8, 3.9)

IL-12/23i (n=1058): 5 (0.5), 3.1 (0.4, 5.8)

IL-17i (n=1163): 8 (0.7), 5.9 (1.8, 9.9)

Apremilast (n=1885): 8 (0.4), 5.2 (1.6, 8.9)

IPTW Cox HRw (95%CI), p-value

IL-1223i: 2.0 (1.3-3.0), <10^-4

IL-17i: 1.9 (1.2-3.0), <10^-3

Apremilast: 1.3 (0.8-2.2), 0.31

(ref: TNF inhibitors)

IPTW Fine-Gray sHRw (95%CI), p-value

IL-1223i: 2.1 (1.5-2.9), <10^-3

IL-17i: 2.3 (1.5-3.0), <10^-3

Apremilast: 1.4 (0.8-2.4), 0.12

(ref: TNF inhibitors)

Article conclusion

Analysis of a large database revealed a small overall number of MACEs, and the risk of MACEs was greater for PsA new users of IL-12/23 and IL-17, not apremilast vs TNF inhibitors.

Subroup and sensitivity analyses

- The results did not differ for patients without skin psoriasis

requiring local treatment.

- Among patients without comorbidities related to CVD, the overall crude incidence was 2.0 (0.0) per 1000 PY. Risk of MACEs was significantly higher (overall P<0.01) for IL-

17 inhibitors (HRw 1.6, 95% CI 1.1, 2.8), but not apremilast

(HRw 0.7, 95% CI 0.4, 1.5), than TNF inhibitors.

- The per-protocol and additional sensitivity analyses results were consistent with those of the main analysis.

Song, 2022

Outcome malignancies

Type of study:

retrospective cohort study

Setting and country:

Korean National Health Insurance database

Funding and conflicts of interest:

- Supported by a grant from the PACEN funded by the Ministry of Health and Welfare, Republic of Korea (grant number: HI19C0481, HC19C0052), and by the Basic Science Research Programme through the NRF funded by the Ministry of Education (NRF-2021 R1A6A1A03038899).

- Y-KS has received research grants from Bristol-Myers Squibb, Eisai, Pfizer and JW Pharmaceutical. Other authors declare no conflict of interest.

Inclusion criteria:

Patients with RA who received their first JAKi or TNFi between 2015-2019.

Exclusion criteria:

- all patients with prescriptions of JAKi or TNFi before the index date to ensure patients were TNFi/JAKi naïve.

- under 18 years of age

- diagnosed with AS, PsA, IBD, JIA

patients

- with prior malignancy in the past 5 years from the index date (hence not in remission)

- with observation periods of less than 6 months

N total at baseline: 4929 patients

Intervention: 1064

Control: 3865

Important prognostic factors²:

age ± SD:

I: 55.7±12.5

C: 54.2±13.3

Sex:

I: % M 17.2

C: % M 20.9

Groups comparable at baseline?

No, even after IPTW, some differences remained.

Describe intervention (treatment/procedure/test):

First JAKi

Describe control (treatment/procedure/test):

First TNFi

Length of follow-up:

Patients were followed up from the index date to the occurrence of malignancy, drug discontinuation, death or the end of the study in December 2019.

Loss-to-follow-up:

Patients who were lost to follow-up were not considered separately because they were censored due to drug discontinuation.

Incomplete outcome data:

n/a

Outcome measures and effect size:

Before IPTW: IR (95%CI); HR (95%CI)

Overall malignancy

JAKi 0.54 (0.26 to 1.14); 0.69 (0.30 to 1.56)

TNFi 0.85 (0.66 to 1.10)

Solid malignancy

JAKi 0.47 (0.21 to 1.04); 0.61 (0.26 to 1.47)

TNFi 0.82 (0.63 to 1.07)

Haematological malignancy

JAKi 0.08 (0.01 to 0.55); 2.41 (0.15 to 37.99)

TNFi 0.03 (0.01 to 0.12)

After IPTW: IR (95%CI); HR (95%CI)

Overall malignancy

JAKi 0.67 (0.48 to 0.94); 0.83 (0.55 to 1.27)

TNFi 0.85 (0.67 to 1.07)

Solid malignancy

JAKi 0.61 (0.43 to 0.87); 0.77 (0.50 to 1.19)

TNFi 0.82 (0.65 to 1.04)

Haematological malignancy

JAKi 0.06 (0.02 to 0.19); 2.86 (0.41 to 20.00)

TNFi 0.02 (0.01 to 0.10)

In the sensitivity and subgroup analyses, there was also no significant difference in the risk of malignancy between the JAKi and TNFi groups.

Article conclusion:

Malignancy risk in Korean patients with RA was not increased with JAKi use compared with TNFi use.

NB: since the observation period was less than 2 years for both groups, it may have been insufficient for malignancies to develop.

NB2: 6% of the total population had used non-TNFis before TNFi/JAKi treatment, but included as covariate.

Song, 2023a

Cardiovascular outcomes (MACE, VTE)

Type of study:

Observational cohort study

Setting/Source and country:

National Health Insurance Service database

Republic of Korea

Funding and conflicts of interest:

This research was supported by a grant of Patient-Centered Clinical Research Coordinating center (PACEN) funded by the Ministry of Health & Welfare, Republic of Korea. This research was also supported by Basic Science Research Program

through the National Research Foundation of Korea (NRF), funded by

the Ministry of Education.

YKS has received research grants from Bristol-Myers Squibb, Eisai, Pfizer, and JW Pharmaceutical. Other authors declare that there are no

conflicts of interest.

Inclusion criteria:

- Prevalent RA = RA diagnostic code + prescription of DMARD

- New user of JAKi or TNFi

- Between 2015 and 2019

Exclusion criteria:

- age <18

- diagnosed with other autoimmune

diseases, such as ankylosing spondylitis, psoriatic arthritis, inflammatory bowel disease and juvenile idiopathic arthritis

- prior history of VTE any time before the index date or a prior history of anticoagulant therapy within 30 days of the index date

- Patients with an observation period of less than six months

N total at baseline: 4178 patients

Intervention: 871 JAKi users

Control: 3307 TNF inhibitor users

Important prognostic factors²:

age ± SD:

I: 54.6 ± 12.3

C: 54.2 ± 13.2

Sex:

I: % M 16.9

C: % M 19.5

Groups comparable at baseline?

No, but IPTW was applied.

After performing sIPTW, 577 JAKi users and 3160 TNF inhibitor

users were included. Almost all variables were balanced, including age, sex, comorbidities and medication use. However, several variables were still unbalanced after sIPTW: seropositivity, Charlson Comorbidity Index score group, varicose vein and obesity.

Describe intervention (treatment/procedure/test):

JAKi: tofacitinib or baricitinib

Describe control (treatment/procedure/test):

TNFi: adalimumab, etanercept, golimumab, or infliximab

Length of follow-up:

- The observation started

from the index date and terminated at the incidence of VTE, discontinuation of JAKi or TNF inhibitor, death, or December 2019.

- 12 weeks of permissible gap to define discontinuation of JAKi or TNF inhibitor, so a gap of less than 12 weeks in addition to the usual drug

interval was not considered drug discontinuation. As-treated analysis was mainly performed for VTE incidence, and intention-to-treat (ITT) analysis was conducted as a sensitivity analysis.

Loss-to-follow-up:

-

Incomplete outcome data:

-

Relevant measures and effect size:

The risk of VTE in RA patients treated with JAKi versus TNFi (after sIPTW)

N patients, n events, person-years, IR(95%CI), HR (95%CI), aHR (95%CI)

VTE

JAKi: 577, 0.4, 688.0, 0.06 (0.00–1.23), 0.18

(0.01–3.47), 0.18 (0.01–3.47)

TNFi: 3160, 20.4, 5384.9, 0.38 (0.25–0.58), ref., ref.

PE

JAKi: 577, 0, 688.1, N/A, N/A, N/A.

TNFi: 3160, 4.1, 5395.2, 0.08 (0.03–0.20)

DVT

JAKi: 577, 0.4, 688.0, 0.06 (0.00–1.23), 0.21

(0.01–4.24), 0.22 (0.01–4.32)

TNFi: 3160, 17.1, 5388.2, 0.32 (0.20–0.51)

Article conclusion:

There is no increased risk of VTE in RA patients treated with JAKis compared with TNF inhibitors in Korea.

In sensitivity analyses, ITT analysis was conducted in two ways: for the total observation period and 1 year. The adjusted HR of VTE

was 0.33 (95% CI 0.06–1.90) in the ITT analysis for the total observation period and 0.40 (95% CI 0.04–3.71) in the ITT analysis for 1 year.

A subgroup analysis according to sex, age, hypertension and dyslipidaemia was conducted. The HR could not be calculated in several subgroups due to the absence of VTE incidence amongst JAKi

users, but there were no statistically significant differences in those subgroups where the HR could be calculated. When tofacitinib users and

TNF inhibitor users were compared, tofacitinib did not increase the risk of VTE compared with TNF inhibitor.

Song, 2023b

Infections (serious, herpes zoster)

Type of study:

Prospective observational cohort study

Setting/Source and country:

academic referral hospital in Korea

Funding and conflicts of interest:

This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea. Tofacitinib registry and this research was sponsored by Pfizer.

YKS has received research grants from Bristol-Myers Squibb, Eisai, Pfizer, and JW Pharmaceutical. JYJ and HJY are employees and shareholders of Pfizer Inc. No potential conflicts of interest relevant to this article are reported. The other authors declare that they have no competing interests.

Inclusion criteria:

All patients with RA who started targeted therapy in our institution were candidates for the cohorts, but those who refused to write an informed consent were excluded. In this study, patients with RA starting tofacitinib between March 2017 and

May 2021 and those starting TNFi between July 2011 and May 2021 were included.

Exclusion criteria:

No informed consent.

N total at baseline: 912

Intervention: 200

Control: 712

Important prognostic factors²:

age ± SD:

I: 53.4 ± 12.2

C: 50.8 ± 13.5

Sex:

I: % M 9.5

C: % M

Groups comparable at baseline?

No, but Differences in the baseline variables were balanced after IPTW

between 200 tofacitinib users and 200 TNFi users.

Describe intervention (treatment/procedure/test):

Tofacitinib

Describe control (treatment/procedure/test):

TNFi: etanercept, infliximab, adalimumab, golimumab

Length of follow-up:

Demographic and clinical information of RA patients were collected at enrolment, and the disease activity was assessed every 6 months according to DAS28 and PROMs.

The observation period commenced from the initiation of tofacitinib or TNFi and continued until the onset of HZ, discontinuation of each agent, or May 2021.

Loss-to-follow-up:

-

Incomplete outcome data:

The medical records of patients were reviewed together to minimize the possibility of missing

information on the study-related data points.

Relevant measures and effect size:

Incidence and risk of Herpes Zoster and serious HZ, tofacitinib vs. TNFi

N cases, IR/100PY, IRR (95%CI), p-value

Incidence and risk of HZ Before IPTW

TOFA: 20, 6.03, 3.27 (1.89–5.65), <0.001

TNFi: 36, 1.85, ref.

Incidence and risk of serious HZ Before IPTW

TOFA: 1, 0.30, 0.99 (0.12–8.20), 0.987

TNFi: 6, 0.31, ref.

Incidence and risk of HZ after IPTW

TOFA: 20, 6.03, 8.33 (3.05–22.76), <0.001

TNFi: 5, 0.77, ref.

Incidence and risk of serious HZ after IPTW

TOFA: 1, 0.30, 3.52 (0.14–93.51), 0.452

TNFi: 1, 0.15, ref.

Incidence and risk of HZ within 12 months of use before IPTW

TOFA: 12, 7.39, 4.59 (1.93–10.88), <0.001

TNFi: 9, 1.61, ref.

Incidence and risk of HZ within 12 months of use after IPTW

TOFA: 12, 7.39, 10.83 (1.58–74.51), 0.015

TNFi: 1, 0.60, ref.

Article conclusion:

Tofacitinib use increased the risk of HZ compared with TNFi in Korean patients with RA, but the rate of serious HZ or permanent discontinuation of tofacitinib due to HZ event was low.

In the sensitivity analysis, the IRR of HZ development within 12 months increased significantly after IPTW to 10.83 (95% CI 1.58–74.51, p = 0.015)

Uchida, 2023

Various relevant outcomes

Type of study:

retrospective cohort study

Setting and country:

Nagasaki University

Hospital, Sasebo Chuo Hospital and Ureshino Medical Center, Japan

Funding and conflicts of interest:

No specific funding was received. No conflicts of interest.

Inclusion criteria:

RA diagnosis

treated with tofacitinib, baricitinib, or TNF inhibitors

informed consent

Exclusion criteria:

None reported.

N total at baseline: 499

Intervention: 296 (TAFA 192, BAR 104)

Control: 203

Important prognostic factors²:

age (IQR):

I: TOFA 67 (58–73)

I: BAR 68 (58–75)

C: TNF 51 (37–61)

Sex:

I: % M TOFA 19.3

I: % M BAR 15.4

C: % M TNF 18.2

Groups comparable at baseline?

No, but IPTW is performed.

Describe intervention (treatment/procedure/test):

The JAK inhibitor–treated patients received either tofacitinib

5mg twice or once daily (in patients with renal impairment)

or baricitinib 2mg (in patients with renal impairment)

or 4mg once daily.

Describe control (treatment/procedure/test):

TNF inhibitors (adalimumab

or etanercept)

Length of follow-up:

Max. March 2013 through December 2020; the period from the initiation of

tofacitinib or baricitinib treatment until the treatment’s discontinuation,

the patient’s transfer to another hospital or death, or the end of the study period, whichever occurred first.

Loss-to-follow-up:

Is when outcomes occur.

Incomplete outcome data:

None reported.

Follow-up time, yrs:

TOFA 2.0 (0.5–3.8)

BAR 0.8 (0.33–1.5)

TNF 1.4 (0.4–3.1)

Outcome measures and effect size:

n(%); IR per 100 PY (95% CI)

Serious infectious diseases other than HZ

JAKi 40 (13.5%); 8.36 (7.79, 8.93)

TOFA 33 (17.2%); 8.68 (8.10, 9.26)

BAR 7 (6.7%); 7.09 (6.57, 7.61)

TNFi 16 (7.9%); 4.09 (3.70–4.49)

Herpes Zoster

JAKi 57 (19.3%); 13.00 (12.29, 13.71)

TOFA 46 (24.0%); 13.37 (12.65, 14.09)

BAR 11 (10.6%); 11.63 (10.96, 12.30)

TNFi 6 (3.0%); 1.48 (1.24–1.72)

MACE

n = 2 (1.0%) in TOFA

0 and 0 in BAR and TNF

Malignancies

n = 11 (3.7%) in JAKi

n = 3 (1.4%) in TNFi

The weighted Cox proportional hazard model revealed the following adjusted hazard ratios

TNFi vs JAKi

Serious infection

0.792 (95% CI: 0.417, 1.502), p=0.475

HZ

0.200 (95% CI: 0.077, 0.524), p=0.001

Malignancy

0.385 (95% CI: 0.095, 1.552), p=0.179

Article conclusions:

The infectious disease IR in RA was comparable between tofacitinib and baricitinib, but the IR for HZ in these treatment groups was high compared with that in the TNF inhibitor treatment group. The malignancy rate in the JAK-inhibitor-treated group was high but not significantly different from that of the general population or that of the TNF-inhibitor–treated group.

Note: baseline data were limited and did not include parameters of a smoking history, rate of vaccination other than the HZ vaccine, or disease activity.

Westermann, 2023

Outcome malignancies

Type of study:

observational cohort study

Source and country:

the Danish Rheumatology Quality Register, the Danish Cancer Registry, the Danish Civil Registration System, the Danish Population Education Register, and the Danish National Patient Register, Denmark

Funding and conflicts of interest:

This work was financially supported by the Danish Rheumatism Association and the Danish Cancer Society. The funders were involved neither in the planning, performance,

nor the submission of the manuscript.

Two of the seven authors have no conflicts of interests. For the other five, disclosures are provided in an extensive statement

Inclusion criteria:

All patients with RA in DANBIO

aged 18 or more and who

initiated treatment with either JAKi or bDMARD’s (index date)

from 1 January 2017 to 31 December 2020.

Exclusion criteria:

Patients with a registered cancer prior to the index date were excluded, except for those with prior

non-melanoma skin cancer (NMSC).

For both groups, patients with previous bDMARD treatments were included.

N total at baseline: 4601

Intervention: 875

Control: 4247

Important prognostic factors²:

age ± SD:

I: 57.8 (13.05)

C: 57.6 (13.83)

Sex:

I: % M 20.2

C: % M 24.9

Groups comparable at baseline?

Not before IPTW, but after all covariates and covariate levels had post-IPTW weighting SMDs <10%, and the majority were <5%.

Describe intervention (treatment/procedure/test):

any available JAKi in Denmark, i.e. tofacitinib

or baricitinib

Describe control (treatment/procedure/test):

interleukin-6 inhibitors (tocilizumab/

sarilumab); B-cell inhibitors via anti-CD20 (rituximab);

T-cell co-stimulation inhibitors via CTLA-4 (abatacept); all

types of TNFi,

including both originators and

biosimilars

Length of follow-up:

Follow-up started on the date upon first registered treatment with JAKi or bDMARD’s after 1 January 2017. Follow-up was performed in a hierarchical ever-treated design, i.e. ‘once exposed always exposed’. However, patients were able to switch from the bDMARD group to the JAKi group, mimicking the usual hierarchical treatment pattern for JAKi-treated

patients according to Danish guidelines. Person years (PYRS) of follow-up and number of cancers were allocated to each JAKi and bDMARD group, respectively, by following

patients from the index date and until date of cancer, death, emigration, initiation of JAKi (bDMARD group censoring only), or 31 December 2020, whichever occurred first.

During follow-up, the JAKi group contributed 1315 PYRS (median 1.48 years; interquartile range 0.98–

1.93 years) and 19 cancers, while the bDMARD group contributed 8597 PYRS (1.98; 1.11–3.09) and 111 cancers.

Outcome measures and effect size:

Malignancy (excl. NMSC)

JAKi IR/1000PY 14.4

bDMARD IR/1000PY 12.9

weighted CSC model:

HR JAKi vs bDMARD: 1.41 (95%CI 0.76, 2.37)

Article conclusion: JAKi treatment in real-world patients with RA was not associated with a statistically significant increased risk of first primary cancer

compared with those who received bDMARD’s.

Note: Analyses stratified by age and by length of follow-up also displayed numerically increased yet statistically non-significant HRs. None of the sensitivity analyses showed any statistically significant

HRs

EULAR

Sepriano, 2023

Yes

Yes

Yes

Yes

Unclear

The risk of bias varies per included study.

Yes

No

Studies were heterogeneous, precluding data pooling, and results are presented descriptively.

Yes

Yes

Author, year

Selection of participants

Was selection of exposed and non-exposed cohorts drawn from the same population?

Exposure

Can we be confident in the assessment of exposure?

Outcome of interest

Can we be confident that the outcome of interest was not present at start of study?

Confounding-assessment

Can we be confident in the assessment of confounding factors? 

Confounding-analysis

Did the study match exposed and unexposed for all variables that are associated with the outcome of interest or did the statistical analysis adjust for these confounding variables?

Assessment of outcome

Can we be confident in the assessment of outcome?

Follow up

Was the follow up of cohorts adequate? In particular, was outcome data complete or imputed?

Co-interventions

Were co-interventions similar between groups?

Overall Risk of bias

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Definitely yes, probably yes, probably no, definitely no

Low, Some concerns, High

Fang, 2022

Various relevant outcomes

Definitely yes

Reason: Participants were selected from a registry

Definitely yes

Reason:

Source National Health Insurance research database.

Definitely yes

for all outcomes, except malignancy

Definitely no

for outcome malignancy.

NB: groups were matched on comorbidities.

Reason: Outcome malignancy could already be present at baseline, registered as comorbidity.

Probably yes

Reason: From high quality database, all available confounding factors were assessed.

Some variables associated with the study outcomes were not in the database.

Definitely yes

Reason: Comprehensive matching with propensity score

Probably yes

Reason: Outcomes came from a national research database.

Probably yes

Reason: FU was performed from the first use of agents of interest until the first occurrence of the individual study outcomes, switch to other agents, or end date study, whichever came first.

Definitely yes

Reason: Additional used medication was balanced between groups

Some concerns

Frisell, 2023

Various relevant outcomes

Definitely yes

Reason: Participants were selected from a registry

Definitely yes

Reason: Source is a longitudinal clinical registry-infrastructure with consistent definitions of treatment cohorts, follow-up and outcomes (ARTIS)

Definitely yes

Reason: Patients with a recent history of an outcome (prior 5 yrs, except for infection where only last year was considered) were excluded from analyses of the same outcome.

Definitely yes

Reason: Source is a longitudinal clinical registry-infrastructure with consistent definitions of treatment cohorts, follow-up and outcomes (ARTIS)

Definitely yes

Reason: Groups were balanced by IPTW taking into account all relevant confounders

Definitely yes

ReasonSource is a longitudinal clinical registry-infrastructure with consistent definitions of treatment cohorts, follow-up and outcomes (ARTIS)

Definitely yes

Reason: Data were complete on treatments, outcomes and most

covariates. Missing covariate data were accounted for by multiple imputation.

Probably no

Reason: Treatment groups seem to differ in concomitant medications, but groups were balanced on these variables.

Some concerns

Hirose, 2022

Various relevant outcomes

Probably no

Reason: Data were derived retrospectively for patients starting treatment before start of study and prospectively for patients starting within the duration of the study. No information how many patients/data were obtained retrospectively and how many prospectively; not clear if different per treatment group

Definitily yes

Reason: Secure record

Probably yes

Reason: incidence and severity of all adverse events were recorded during the study period. Co-morbidities (which could be the same as safety outcomes) were not taken into account in the article, but extensive lab measurements were taken into account.

Probably yes

Reason: Adjustment for most plausible prognostic variables obtained from medical records (retrospectively or prospectively).

Probably yes

Reason: inverse probability of treatment

weighting (IPTW) based on a propensity score that reduces the selection bias to a minimum and adjusts for confounding

factors between binary treatment groups was used

Probably yes

Reason: Partly from medical records, partly from prospective data collection.

The common terminology criteria for adverse

events of the National Cancer Institute (v.5.0) were used

to describe and grade adverse events and laboratory

abnormalities.

Probably yes

Reason: The last observation carried forward

method was used for patients who discontinued treatment before week 52 to include all patients in the analysis.

No information: Unclear what they did with respect to data from patients who withdrew their consent for data-usage.

Definitely yes

Reason: Additional used medication was balanced between groups

Some concerns

Hoisnard, 2022

Cardiovascular outcomes (MACE, VTE)

Definitely yes

Reason: Participants were selected from the same registry over the same time period.

Definitely yes

Reason: secure record, identification in registry with ATC classification codes.

Probably yes

Reason: Presence of outcome at baseline was not an exclusion criterium, but also not likely to coincide with index date and risks associated with outcome were identified and taken into account in analyses.

Probably yes

Reason: All relevant available covariates collected from the French national health data

system.Some variables included in the propensity score relied on proxies. No information on missing covariate data.

Probably yes

Reason: Groups were balanced by IPTW taking into account all available relevant confounders. However, confounders such as disease activity data were not available.

Definitely yes

Reason: Identification of outcomes by ICD-10 codes in national database.

Probably yes

Reason:

Patients were followed up to the occurrence of each event (MACE and VTE), death from any cause, exposure discontinuation

or study end, whichever came first.

Definitely no

Reason: Groups differed in concomitant csDMARD use, but analyses were corrected / groups balanced on these variables.

Some concerns

Huss, 2023

OutcomeMalignancies

Probably no

Reason: Same database sources, same locations, but shorter time frame for intervention due to 1 year later introduction of JAKi

Definitely yes

Reason: Secure records

Definitely yes

Reason: Patients with previous cancer diagnosis (outcome) were excluded

Probalby yes

Reason: All plausible confounding variables were recorded, came from various national registries

Probably yes

Reason: adjustment for most plausible confounding variables

Definitely yes

Reason: medical record and record linkage

Probably yes

Reason: FU from every treatment initiation until the occurrence of the outcome, death, emigration from

Sweden or end of the study period

Definitivily no

Reason: % use of MTX and oral steroids seems to vary per group, but no comparative analyses are presented. However, analyses were corrected for confounding of co-interventions.

Some concerns

Jeong, 2022

Infections (serious, herpes zoster)

Definitely yes

Reason: Participants were selected from the same registry over the same time period.

Definitily yes

Reason: Secure record; high quality national registry

Definitely yes

Reason: Patients with (recent or rapidly recurring) herpes zoster (outcome) were excluded

Probalby yes

Reason: All plausible confounding variables were recorded, coming from a good quality national database.

No information: article mentions that patients with missing data, such as covariates, were excluded, but they don’t mention anything about subsequently excluded patients; all originally included patients are in analyses.

Probably yes

Reason: adjustment for most plausible confounding variables

Probably yes

Reason: Outcomes came from a national research database.

Probably yes

Reason: FU until until drug failure, development of herpes zoster (outcome), or study end.

Probably no

Reason: Groups differ, but analyses were corrected for these variables.

Some concerns

Khosrow-Khavar, 2022

Cardiovascular outcomes (MACE, VTE)

Definititely yes

Reason: Participants were selected from the same registry.

Definitily yes

Reason: Secure record; high quality national registry

Probably yes

Reason: assessed 76 potential confounders from various national databases, including medical history.

Definitely yes

Reason:

Propensity score fine stratification achieved excellent covariate balance with standardized differences close to zero across all covariates

Definitely yes

Reason: assessed 76 potential confounders from various national databases.

Probably yes

Reason: Outcomes came from various national databases.

Probably yes

Reason: patients were followed from treatment initiation for study outcomes until treatment discontinuation or switch, insurance disenrollment, death, or end of the study period, whichever occurred first.

Probably no

Reason: Groups differ, but analyses were corrected for this.

Some concerns

Min, 2023

Various relevant outcomes

Definititely yes

Reason: Participants were selected from the same registry.

Definititely yes

Reason: Secure records

Definitely yes

Reason: patients with a main ICD-10 diagnosis code for AMI, stroke, VTE, ATE, or cancer 12 months prior to the initiation of TNFi or JAKi were excluded.

Probably yes

Reason: Most important covariates as recorded in health insurance database

Probably yes

Reason: Authors did adjust analyses for all the covariates that they did collect. No information: Not clear what was done with missing covariate information.

Definitely yes

Reason: National database ICD-10 codes were used

Probably yes

Reason: FU until outcome occurred, switch in medication type, or end of study.

Definitely no

Reason: Glucocorticoid and NSAID us differed between groups.

However, MTX was used in both groups (inclusion criteria) and analyses were corrected for confounding variables.

Some concerns

Mok, 2023

Various relevant outcomes

Definitely yes

Reason: Participants were selected from the same registry

Definitely yes

Reason: Secure records

Probably yes

Reason: Not entirely clear, but history of outcome diseases was recorded and taken into account in analyses.

Probably yes

Reason: Analyses were adjusted for important covariates as recorded in a registry; these differed per outcome.

Probably yes

Reason: Analyses were adjusted for relevant covariates in which groups differed. However, analyses were not always adjusted for all covariates; no information why.

Definitely yes

Reason: Regional registry used and verification by review of individual medical records by

two independent research assistants.

Probably yes

Reason: FU until outcome occurred, start of a new cours, or end of study.

Definitely no

Reason: Between group differences in MTX, LEF, HCQ, Glucocorticoids, sulphasalizine. At least the analyses for outcome Infections were adjusted for this.

Some concerns

Molander, 2022

Cardiovascular outcomes (MACE, VTE)

Definitely yes

Reason: Participants were selected from the same registry over the same time period.

Definitily yes

Reason: Secure record; high quality national registry

Definitely yes

Reason: Subjects with a VTE (outcome) registered during the year prior to start of follow-up

were excluded.

Probably yes

Reason: Analyses were adjusted for important covariates as recorded in a registry

Probably yes

Reason: adjustment for most plausible confounding variables

Definitely yes

Reason: information from various registries was combined to define an outcome.

Probably yes

Reason: follow-up

ended at 60 days after discontinuation of the DMARD treatment in question, first VTE event, death, emigration or end of study period, whichever came first.

Definitely no

Reason: Compared with TNFi, patients

starting other b/tsDMARD’s were generally slightly older, had

longer-lasting

RA, more comorbidities and a higher level of

healthcare consumption. However, analyses were adjusted for this.

Some concerns

Pina Vegas, 2022

Cardiovascular outcomes (MACE, VTE)

Definitely yes

Reason: Participants were selected from the same registry over the same time period.

Definitily yes

Reason: Secure record; high quality national registry

Definitely yes

Reason: patients with a history of acute myocardial infarction,

unstable angina, chronic ischaemic heart disease,

ischaemic stoke or transient ischaemic attack were excluded.

identified within 5 years before the index date

Definitely yes

Reason: Analyses were adjusted for important covariates as recorded in a registry. IPTW was performed.

Definitely yes

Reason: Comprehensive matching with propensity score.

Probably yes

Reason: Events were identified by a

hospital discharge diagnosis - ICD-10 codes - with a previously validated algorithm.

Probably yes

Reason: Patients were followed

up to the MACE event, death from any-cause, systemic treatment switch, lost to follow-up (defined by the absence of any reimbursement for 12 consecutive

months) or study end, whichever came first.

Definitely no

Reason: groups differed on various covariates. However, after IPTW, a pseudo-cohort with a similar distribution of between the different treatment classes was obtained.

Some concerns

Song, 2022

OutcomeMalignancies

Definitely yes

Reason: Participants were selected from the same registry

Definitely yes

Reason: Secure records

Definitely yes

Reason: Outcome was cancer and cancer had to be in remission at start of treatment (cancer in last 5 years meant exclusion)

Probably yes

Reason: Most relevant covariates, but maybe not all/those available were taken into account, as recorded in health insurance database

Probably yes

Reason: Comprehensive matching with propensity score. After IPTW, authors report still unbalanced

covariates including the year of initiating JAKi or TNFi treatment, seropositivity, cerebrovascular disease, previous csDMARD use, and concomitant MTX and oral corticosteroid use.

Definitely yes

Reason: linkage of two records

Definitely yes

Reason: Patients were followed up from the index date to the occurrence of malignancy, drug discontinuation, death or the end of the study in December 2019.

Patients who were lost to follow-up were not considered separately because they were censored due to drug discontinuation.

Definitely no

Reason: Even after IPTW, differences remained on concomitant MTX and oral corticosteroid use.

Some concerns

Song, 2023a

Cardiovascular outcomes (MACE, VTE)

Definitely yes

Reason: Participants were selected from the same registry

Definitely yes

Reason: Secure records

Definitely yes

Reason: those with a prior history of VTE

or a prior history of anticoagulant therapy within 30 days of the

index date were excluded.

Probably yes

Reason: Most relevant covariates, but maybe not all/those available were taken into account, as recorded in health insurance database

Probably yes

Reason: Comprehensive matching with propensity score. However, several variables were

still unbalanced after sIPTW.

Probably yes

Reason: defined by ICD10 code and use of anticoagulant therapy within 30 days from VTE diagnosis

Probably yes

Reason: The observation started

from the index date and terminated at the incidence of VTE, discontinuation

of JAKi or TNF inhibitor, death, or study end.

Definitely no

Reason: Differences before IPTW, but not after on medication use.

Some concerns

Song, 2023b

Infections (serious, herpes zoster)

Probably yes

Reason:

Same hospital, all at start target medication

Definitely yes

Reason: Secure records, medical record

Probably no:

Reason: no exclusion criteria for (recent or rapidly recurring) herpes zoster (outcome)

Probabably yes

Reason: Most relevant covariates were recorded.

Definitely yes

Reason: Comprehensive matching with propensity score. Groups balanced after IPTW.

Probably yes

Reason: based on HZ diagnose in medical record and reporting as AE or SAE. Medical records reviewed by two people together.

Probably yes:

Reason: The observation period commenced from the initiation of tofacitinib or TNFi and continued until the onset of HZ, discontinuation of each agent, or study end.

Definitely no

Reason: Differences before IPTW, but not after on medication use.

Some concerns

Uchida, 2023

Various relevant outcomes

Probably yes

Reason: participants from same points of care over the same time frame

Definitely yes

Reason: Medical record

Probably yes

Reason: For the outcome of malignancies they excluded recurrent ones or occurrence within one month after treatment start.

Probably no

Reason: Most relevant covariates recorded (medical record) at initiation of treatment, but as the authors note, not all possible important confounders were taken into account

Definitely yes

Reason: Comprehensive matching with propensity score

Definitely yes

Reason: Outcomes were determined by treating physician(s) according to described requirements.

Definitely yes

Reason: Patients were followed up from the initiation of

tofacitinib or baricitinib treatment until the treatment’s discontinuation,

the patient’s transfer to another hospital or

death, or the end of the study period, whichever occurred

first.

No information

Reason: Groups seem to differ on concomitant treatments before IPTW. Not clear if they were balanced after IPTW.

Some concerns

Westermann, 2023

OutcomeMalignancies

Definitely yes

Reason: Participants were selected from the same registry

Definitely yes

Reason: high quality databases

Definitely yes

Reason: patients with a registered cancer (outcome) prior to index date were excluded.

Probably yes

Reason: Indicators of all possible relevant confounders were collected from high quality national databases.

NB: Strangely enough, concomitant and/or previous MTX use is nowhere mentioned in te article.

Definitely yes

Reason: Comprehensive matching with propensity score

Definitely yes

Reason: All cancer diagnoses were collected from the DCR. Cancer diagnoses registered in DCR have been estimated to be highly valid and complete.

Definitely yes

Reason: FU until date of: cancer, death, emigration, initiation of JAKi (bDMARD group censoring

only), or 31 December 2020, whichever occurred first. Death was considered a competing risk

due to its preclusion of cancer occurrence. We calculated

95% bootstrap CIs (95% CI) with 500 iterations for all

weighted models

Probably no

Reason: Not all possible concomitant medication per group provided (e.g., MTX use is missing)

Some concerns