Interactive modules by Olga Goodman, MD.Copyright protected · Read termsClose terms

All RheumaView™ interactive modules, quizzes and educational cases are original works developed by Olga Goodman, MD.

Without the author’s prior written permission, you may not copy or republish these modules or their components; extract, reuse or modify their original software code (“engines”); or use that code to build your own modules, quizzes or other products.

You may not claim these works as your own or remove or replace the author’s name or copyright notices. When quoting material, credit Olga Goodman, MD, identify the source and include a direct link to the original.

Free or public access is provided for learning and does not grant permission to copy, adapt or redistribute the materials. Unauthorized use of protected material may constitute copyright infringement. Exceptions and limitations provided by applicable law remain unaffected.

Sharing direct links to the modules is welcome.

© 2026 Olga Goodman, MD. All rights reserved.

RHEUMATOLOGY IN PERSPECTIVE · RHEUMAVIEW

SJÖGREN’S · Clinical trials

Sjögren clinical trials: results and lessons

Positive, negative and incomplete evidence, read against the actual question · Selected evidence updated 11 November 2026

This is a clinical evidence archive, not a league table of drugs. The studies below differ in disease activity, serology, residual gland function, background therapy, endpoints and duration. Numerical results are attributed to the report actually available: a peer-reviewed article or abstract, a conference-related sponsor presentation, or a topline announcement. Trial doses describe research exposure; they are not prescribing instructions.

Read the experiment before the headline

A positive result answers a defined question in a defined population. Record the inclusion criteria, randomized groups, comparator, background care, outcome, time point and analysis population. Then distinguish change within the active arm from the extra improvement over the comparator.

Also ask who was excluded, how rescue treatment and missing data were handled, whether the primary endpoint was prespecified, and whether secondary analyses were protected against multiple testing. Randomization supports causal inference; an uncontrolled before–after improvement does not isolate a treatment effect. A press release cannot supply details it has not disclosed.

Back to contents ↑

The outcomes are not interchangeable
Outcome familyWhat it is intended to captureWhat it cannot establish alone
ESSDAIClinician-assessed systemic activity across domainsRelief of every symptom, permanent damage reversal or gland regeneration
ClinESSDAIClinical systemic activity without the biological domainA direct comparison with an ESSDAI effect from another population
ESSPRI and other symptom measuresPatient-reported burdenAttribution of symptoms to immune inflammation
Salivary flow, ocular testsSpecific aspects of gland or surface functionPrevention of all dental/ocular complications or overall disease control
CRESS, candidate STARComposite response across several domainsUniversal clinical validation for every future use or a guarantee that each component improved

Different endpoint choices help explain why a drug can change laboratory measures without succeeding clinically, or improve systemic activity while leaving symptoms largely unchanged. These instruments are discussed conceptually here; the site does not reproduce a scoring instrument or offer an unvalidated calculator. [R20] [R28] [R33] [R34] [R35]

Back to contents ↑

JOQUER: hydroxychloroquine and the core symptom triad

Design: France, 120 participants; HCQ 400 mg/day versus placebo for 24 weeks, then HCQ for both groups. The primary response required at least 30% improvement in two of three symptom scales: dryness, pain and fatigue. [R20]

Result: 17.9% (10/56) versus 17.2% (11/64); odds ratio 1.01, 95% CI 0.37–2.78, p=0.98. Serious adverse events during the blinded period numbered two versus three. The primary endpoint was negative.

Meaning: this trial does not support routine HCQ for the combined symptom target in an unselected similar population. It does not comprehensively test all inflammatory arthritis, skin or overlap indications. Later subgroup analyses can suggest a new trial hypothesis, but cannot retrospectively make the original primary endpoint positive.

Back to contents ↑

TRACTISS: rituximab did not establish broad symptom benefit

Design: 133 anti-Ro-positive UK participants with fatigue and oral dryness; rituximab 1,000 mg at weeks 0, 2, 24 and 26 versus placebo, with infusion co-medication. Primary endpoint: at least 30% improvement in fatigue or oral dryness at week 48. [R21]

Result: imputed response rates 39.8% versus 36.8%; adjusted odds ratio 1.13, 95% CI 0.50–2.55. Most outcomes did not differ; unstimulated salivary flow was an exception. Serious adverse events numbered ten in each arm.

Meaning: a biological rationale and B-cell depletion are insufficient evidence of symptom efficacy. Conversely, this population and endpoint do not settle treatment of rare organ-threatening vasculitis or neuropathy. An isolated positive secondary measure must remain separate from a negative primary result.

Back to contents ↑

Abatacept: a plausible target, negative confirmatory trial

The multicentre phase III trial randomized 187 participants to subcutaneous abatacept 125 mg weekly or placebo. At day 169, ESSDAI change was −3.2 with abatacept and −3.7 with placebo, p=0.442. Key symptom and stimulated salivary outcomes also failed to show superiority. [R22]

This illustrates why uncontrolled improvements and changes in immune biomarkers cannot substitute for a blinded comparator. It also cautions against reasoning from a drug’s efficacy in rheumatoid arthritis to assumed efficacy in Sjögren disease. The finding concerns the tested regimen and population; it does not prove that all T-cell-directed approaches must fail.

Back to contents ↑

BELISS: a signal, not a placebo-adjusted response rate

BELISS was an open-label phase II study of belimumab in 30 participants. Eighteen, or 60%, met the chosen composite response at week 28. The absence of a randomized control group means this is an observed responder proportion, not the percentage whose improvement was caused by belimumab. [R47]

The composite included symptoms, physician-assessed systemic activity and B-cell biomarkers. A response could therefore arise through different components. The study supplied a rationale for further testing, but cannot be directly ranked against the response percentages of placebo-controlled trials using other endpoints.

Back to contents ↑

Sequential belimumab–rituximab: target engagement versus confirmation

A four-arm phase II study randomized 86 patients to placebo, belimumab, rituximab or the sequence of both; 60 completed treatment and follow-up to week 68. Combination treatment produced deeper glandular B-cell depletion. Mean ESSDAI fell from 11.0 to 5.0 in the combination arm and from 10.4 to 8.6 with placebo. [R25]

These are informative descriptive values, not a definitive comparative efficacy claim from a large confirmatory trial. The study emphasized safety and biological effects, and the small arms and incomplete follow-up limit precision. Infections were common adverse events; similar event rates in a small trial do not exclude uncommon or delayed harms.

Back to contents ↑

RepurpSS-I: combining existing medicines

The single-centre phase 2a trial randomized 29 adults with ESSDAI ≥5 and a positive gland biopsy: 21 to leflunomide 20 mg plus HCQ 400 mg daily, eight to placebo, for 24 weeks. The adjusted between-group ESSDAI difference was −4.35 points, 95% CI −7.45 to −1.25, p=0.0078. [R23]

Gastrointestinal discomfort occurred in 52% versus 25%, and transient ALT elevations in 48% versus 13%; small denominators make percentages unstable. One placebo participant requiring high-dose prednisone was excluded from the primary analysis. This positive small study justified replication, rather than automatic adoption for every symptom phenotype.

Back to contents ↑

RepurpSS-II and NECESSITY: populations and endpoints

The 2026 phase 2b trial randomized 46 adults with ESSDAI ≥5 to the same combination (n=21) or placebo (n=25). At 24 weeks, the adjusted between-group ESSDAI difference was −4.135, 95% CI −6.558 to −1.709, p=0.0012. Gastrointestinal discomfort occurred in 6/21 versus 4/25; one serious event, myocardial infarction, was judged probably unrelated. [R24]

Replication strengthens the systemic-activity signal. It remains a small, single-centre study, not proof of benefit in high-symptom/low-activity disease, children or pregnancy. Trial exclusions and the reproductive implications of leflunomide matter. The published abstract provides the adjusted contrast; arm-level change estimates are not reconstructed here from the difference.

NECESSITY: a broader combination trial

ACR 2026 LB15 reports 211 randomized participants across two cohorts. At week 24, pooled STAR responses were 24.3% with placebo, 54.4% with HCQ–MMF and 52.4% with HCQ–leflunomide (both p<0.001). Differences were significant in the high-symptom/ESSDAI <5 cohort; the ESSDAI ≥5 cohort comparisons were not significant. CRESS differences were also not significant. Serious adverse events were reported in 4.3%, 11.1% and 2.9%, respectively. [R52]

This multinational result extends the evidence beyond RepurpSS, while remaining a conference abstract. The pooled primary result was positive; the subgroup and alternative-endpoint findings qualify its interpretation rather than proving absence of benefit or a general treatment rule.

Back to contents ↑

Ianalumab phase 2b: understand the primary statistical question

In 190 participants with ESSDAI ≥6 and ESSPRI ≥5, monthly subcutaneous ianalumab 5, 50 or 300 mg was compared with placebo over 24 weeks. The primary dose-response objective was met in four of five models. However, the 300 mg versus placebo contrast was −1.92 ESSDAI points, 95% CI −4.15 to 0.32, p=0.092. [R26]

Both facts belong in the same account. A positive model-based dose-response test does not mean every individual dose comparison was significant. There were treatment-related serious events in both placebo and active groups; the trial size and duration do not establish rare-event safety. This study supported further development of BAFF-R blockade with B-cell depletion.

Back to contents ↑

NEPTUNUS: phase III success with modest average separation

The October 2026 NEJM article reports 275 participants randomized in NEPTUNUS-1 and 504 in NEPTUNUS-2. Monthly ianalumab 300 mg improved mean ESSDAI versus placebo at week 48: −1.3 points (95% CI −2.6 to 0.0; p=0.0496) and −1.0 (−2.0 to 0.0; p=0.04), respectively. [R48]

The table retains the earlier sponsor presentation’s arm-level changes and reported precision. Registry enrollment of 506 for NEPTUNUS-2 is a different reported denominator from the 504 randomized in the article. [R41] [R46]

ComparisonActive changePlacebo changeAdjusted difference, 95% CI; p
NEPTUNUS-1, monthly 300 mg−6.4−5.1−1.3 (−2.6 to 0.0); 0.0496
NEPTUNUS-2, monthly 300 mg−6.5−5.5−1.0 (−2.0 to 0.0); 0.041
NEPTUNUS-2, every 3 months−6.0−5.5−0.5 (−1.5 to 0.5); 0.3413

Confidence interval limits are rounded. NEJM reports no significant difference in ESSDAI response (≥5-point reduction) or SSSD/ESSPRI symptom change. Serious adverse events were similar; nonserious local and systemic injection reactions were more common with ianalumab. Mean systemic improvement therefore does not establish broad symptom relief. A within-person response threshold is not interchangeable with a between-group mean difference. [R48]

ACR 2026 evidence update · 14 November 2026

Pooled safety through week 52

The ACR 2026 pooled NEPTUNUS safety analysis included 305 participants receiving monthly ianalumab and 307 receiving placebo. Local injection-site reactions occurred in 28.5% versus 3.3%, and reported neutropenia in 4.3% versus 1.6%. No febrile neutropenia was reported. Serious adverse events occurred in 6.9% versus 9.8%; this descriptive difference does not establish safety superiority. The analysis adds useful event-specific detail to the efficacy discussion, while leaving uncommon and longer-term risks uncertain. It should not be conflated with regulatory approval or with proof of broad improvement in dryness, pain and fatigue.

Sources: Pooled ianalumab safety through week 52

Longer follow-up: encouraging durability, a different comparison

ACR LB14 reports that 610 of 666 eligible core-study completers entered the extension and 551 reached week 108. With continuous monthly treatment, mean ESSDAI change was −7.5 and low disease activity rose from 57.1% at week 48 to 65.1% at week 108. No new safety findings were reported through week 204. [R50]

The reported groups continued treatment or switched from placebo; there was no continuing placebo comparator. The extension includes a blinded two-arm period followed by an open-label period. Its descriptive follow-up cannot replace the randomized symptom comparison. [R56]

Back to contents ↑

Dazodalibep phase II: two questions, two populations

The randomized phase II trial used two separate populations. Population 1 (n=74) had moderate-to-severe systemic activity; day-169 ESSDAI changes were −6.3 with dazodalibep and −4.1 with placebo, p=0.0167. Population 2 (n=109) had high symptom burden with limited systemic activity; ESSPRI changes were −1.8 and −0.5, p=0.0002. Simple subtraction gives separations of −2.2 and −1.3 points, respectively. [R28]

Some participants had associated RA or SLE, unlike many primary-Sjögren-only trials. That broadens one aspect of applicability while complicating cross-trial comparisons. Respiratory infections, headache, diarrhea and other events were reported. The crossover extension and early efficacy findings do not establish long-term comparative safety.

Back to contents ↑

OASIZ-301: systemic activity and dryness at week 48

ACR 2026 LB22 reports 648 randomized participants: dazodalibep 1,500 mg IV every four weeks (n=217), 3,000 mg every 12 weeks (n=217), or placebo (n=214). Baseline mean ESSDAI was 10.2. At week 48, ESSDAI changes were −5.6, −5.2 and −3.8: placebo-adjusted differences −1.8 (adjusted p<0.001) and −1.4 (p=0.010). [R49] [R57]

The every-four-week regimen also improved dryness versus placebo: ESSPRI dryness difference −1.17 and DASPRI dryness −0.93 (both adjusted p<0.001). These findings do not establish separate fatigue or pain benefits. The abstract reports similar adverse-event, serious-event and discontinuation rates across groups. [R49]

Use 648 randomized when describing these results. Earlier releases described approximately 621, while the trial enrollment field reports 651; the available sources do not explain the difference. Enrollment fields and randomized analysis populations should retain their labels. [R58]

Clinical meaning: systemic activity and dryness improved in the studied active-disease population. These are conference/sponsor reports, not a full peer-reviewed phase III article or an approved indication. OASIZ-303 separately tests high symptoms with low systemic activity and remains ongoing; OASIZ-304 evaluates longer-term safety. [R57]

Back to contents ↑

Telitacicept phase II: dose is not a simple ladder

Forty-two anti-SSA-positive adults with ESSDAI ≥5 were assigned equally to weekly telitacicept 160 mg, 240 mg or placebo for 24 weeks. The 160 mg contrast was −4.3 ESSDAI points, 95% CI −7.0 to −1.6, p=0.002; the 240 mg contrast was −2.7, 95% CI −5.6 to 0.1, p=0.056. No serious adverse events occurred in active-treatment arms. [R30]

This small trial does not establish that a higher dose is intrinsically worse, or that uncommon harms are absent. It shows why dose selection needs more than a mechanistic assumption. Stable background HCQ was permitted under specified conditions; recent other therapies were restricted, another reason not to compare raw effect sizes with unrelated trials.

Back to contents ↑

Telitacicept China phase III: substantial signal, reporting level matters

The sponsor report describes 381 participants randomized equally to 160 mg, 80 mg or placebo weekly, with standard therapy. At week 24, ESSDAI change was −4.4, −3.0 and −0.6, respectively; both active comparisons had p<0.0001. Reported ≥3-point ESSDAI response was 71.8% for 160 mg versus 19.3% for placebo. [R31] [R32]

The absolute responder difference is 52.5 percentage points, calculated from those reported rates. It should not be turned into a cross-trial superiority claim: eligibility, background treatment, placebo response and reporting completeness differ. The 2026 Chinese approval disclosure is a separate regulatory event. A full benefit–harm appraisal still needs the complete report, denominators and detailed safety tables; “no new safety signal” is not a quantified risk estimate.

Back to contents ↑

DAHLIAS: nipocalimab and clinical systemic activity

DAHLIAS randomized 163 anti-Ro-positive participants with ClinESSDAI ≥6 to IV nipocalimab 5 mg/kg, 15 mg/kg or placebo every two weeks. At week 24, the 15 mg/kg contrast was −2.65 ClinESSDAI points, 90% CI −4.03 to −1.28, p=0.0018. The 5 mg/kg contrast was −0.34, 90% CI −1.71 to 1.03, p=0.68. [R33]

The confidence level must not be mislabeled 95%. The endpoint excludes the biological domain, relevant for a drug that lowers circulating IgG. The primary analysis treated post-discontinuation observations as missing. Similar adverse-event rates across these small groups do not settle long-term infection risk or use in seronegative disease. DAFFODIL is the separate phase III confirmation program. [R41]

Back to contents ↑

Efgartigimod: RHO signal followed by UNITY futility

RHO randomized 34 participants, 23 to weekly IV efgartigimod 10 mg/kg and 11 to placebo; 31 entered the efficacy analysis. At week 24, CRESS response was 45.5% versus 11.1%, a reported 34.4-point difference. Candidate STAR response was 54.5% versus 33.3%. No formal statistical hypothesis was tested. [R34]

Treatment-emergent adverse events occurred in 87.0% versus 63.6%, all grade 1 or 2. The denominators behind efficacy percentages differ from all randomized participants. This is a small proof-of-concept signal, not a statistically confirmed phase III success, and its composite endpoint is not interchangeable with DAHLIAS ClinESSDAI. The FcRn mechanism alone does not justify ranking the two agents.

UNITY: phase III stopped for futility

On 8 October 2026, argenx stopped the subcutaneous efgartigimod UNITY phase III study after an independent committee’s interim review found it unlikely to meet the week-48 ClinESSDAI primary endpoint. No new safety signal was reported. The early RHO signal remains part of the evidence history; UNITY did not confirm it in the phase III setting. [R51]

Back to contents ↑

TWINSS: iscalimab’s systemic and symptom findings differ

TWINSS enrolled 173 participants in a systemic-active dose-ranging cohort and 100 in a low-systemic-activity symptom cohort. At week 24, a prespecified dose-response model was positive. In the systemic cohort, 150 and 600 mg contrasts were −3.0 (95% CI −4.9 to −1.1) and −2.9 (−4.9 to −1.0) ESSDAI points. [R35]

In the symptom cohort, 600 mg did not significantly improve ESSPRI: difference −0.57, 95% CI −1.30 to 0.15, p=0.12. Serious events occurred in both cohorts, with small dose-group denominators. A general description of “benefit in two populations” would conceal the different strength of the two results.

Back to contents ↑

Intracellular pathways: preserve both positive and negative studies

The 73-participant remibrutinib LOUiSSE trial found an ESSDAI advantage, reported contrast −2.86, p=0.003, but not an ESSPRI advantage (0.17, p=0.663). A biological effect and systemic improvement did not automatically translate into better symptoms. The study’s terminated registry status reflects the decision not to undertake its planned second part and to pursue dose evaluation separately, not a reported safety failure. [R36] [R41]

In contrast, the phase II study of filgotinib, lanraplenib and tirabrutinib randomized 152 participants; 150 received treatment. Week-12 protocol-defined response rates were 43.3%, 42.3% and 34.7%, versus 26.7% with placebo; none was statistically significant, and key secondary endpoints were negative. These were different molecules targeting JAK1, SYK and BTK. A negative trial should not be hidden, but neither can it establish that every future inhibitor in those pathways will fail. [R37] [R41]

Back to contents ↑

Low-dose IL-2: an immune-regulation hypothesis with a small RCT

A single-centre trial randomized 60 women to low-dose IL-2 or placebo for 12 weeks with follow-up to week 24. At that point, 20/30 (66.7%) versus 8/30 (26.7%) achieved a ≥3-point ESSDAI improvement, p=0.004. No severe adverse events were reported; infections were less frequent in the IL-2 arm. [R38]

The small sample and single-centre population limit transportability and safety precision. Expansion of regulatory immune cells supports target engagement; it is not a validated individual predictor of benefit. The infection observation is insufficient to recommend IL-2 as infection prevention. Larger independent trials and durable outcomes remain important.

Back to contents ↑

Local dryness trials: match the scale of the conclusion

A paired-eye diquafosol pilot included 17 people with Sjögren dry eye and examined the effect of a single drop on tear volume. It addresses a local, short-term physiological question. It does not establish sustained systemic benefit, long-term prevention of corneal injury or restoration of the whole gland. [R39]

Local studies deserve the same methodological care as biologic trials: the unit of randomization may be an eye rather than a person; eyes are correlated; short-term secretion and patient comfort may disagree. General dry-eye cohorts, including studies that exclude Sjögren disease, provide adjacent evidence and should be labeled accordingly. The glandular-care chapter develops practical treatment choices.

Back to contents ↑

From a trial result to an individual decision

First decide whether the patient resembles the eligible population and whether the outcome addresses their problem. Next examine the comparator, absolute difference, uncertainty, adverse events and treatment burden. Finally check indication-specific authorization and access in the relevant country.

Subgroup findings are most persuasive when prespecified, supported by an interaction test and replicated prospectively. A biomarker associated with response is not automatically a clinically validated treatment-selection test. Open-label extensions can describe durability among continuing participants, but do not recreate the randomized comparison after crossover or selective withdrawal.

The archive’s unresolved questions are explicit: detailed OASIZ 301 data, complete late-stage benefit–harm reports, long-term organ protection, outcomes in underrepresented populations and whether improvements persist after withdrawal. The pipeline chapter shows which registered studies are positioned to address some of these gaps.

Back to contents ↑

Sources

  1. [R20] Gottenberg et al. Hydroxychloroquine: JOQUER randomized trial (2014)
  2. [R21] Bowman et al. Rituximab: TRACTISS randomized trial (2017)
  3. [R22] Baer et al. Abatacept phase III randomized trial (2021)
  4. [R23] van der Heijden et al. RepurpSS-I (2020)
  5. [R24] Wong et al. RepurpSS-II (2026)
  6. [R25] Mariette et al. Sequential belimumab and rituximab (2022)
  7. [R26] Bowman et al. Ianalumab phase 2b dose-finding trial (2022)
  8. [R28] St Clair et al. Dazodalibep phase II: two populations (2024)
  9. [R29] Amgen: OASIZ 301 positive topline announcement (22 September 2026)
  10. [R30] Xu et al. Telitacicept phase II (2023 online; 2024 issue)
  11. [R31] Vor Bio/RemeGen: China telitacicept phase III ACR data (14 October 2025)
  12. [R32] RemeGen statutory disclosure: telitacicept Sjögren indication approval (9 June 2026)
  13. [R33] DAHLIAS: nipocalimab phase II (Lancet 2025)
  14. [R34] Peene et al. Efgartigimod RHO proof-of-concept trial (2026)
  15. [R35] Fisher et al. Iscalimab TWINSS phase 2b (2024)
  16. [R36] Dörner et al. Remibrutinib LOUiSSE phase II (2024)
  17. [R37] Price et al. Filgotinib, lanraplenib and tirabrutinib phase II (2022)
  18. [R38] He et al. Low-dose interleukin-2 randomized trial (2022)
  19. [R39] Yokoi et al. Diquafosol tear-volume pilot in Sjögren dry eye (2016)
  20. [R41] ClinicalTrials.gov: named late-stage Sjögren programs
    Additional source recordshttps://clinicaltrials.gov/study/NCT05349214https://clinicaltrials.gov/study/NCT06104124https://clinicaltrials.gov/study/NCT06245408https://clinicaltrials.gov/study/NCT05673993https://clinicaltrials.gov/study/NCT07404865https://clinicaltrials.gov/study/NCT04968912https://clinicaltrials.gov/study/NCT06741969https://clinicaltrials.gov/study/NCT05817669https://clinicaltrials.gov/study/NCT07760818https://clinicaltrials.gov/study/NCT03905525https://clinicaltrials.gov/study/NCT04035668https://clinicaltrials.gov/study/NCT03100942https://clinicaltrials.gov/study/NCT02446249
  21. [R46] Novartis Q3 2025 results: NEPTUNUS numerical table, slide 17
  22. [R47] Mariette et al. BELISS open-label phase II study (2015)
  23. [R48] Mariette et al. Ianalumab: NEPTUNUS-1 and NEPTUNUS-2, NEJM (21 October 2026)
  24. [R49] Noaiseh et al. OASIZ-301, ACR 2026 late-breaking abstract LB22
  25. [R50] Dörner et al. NEPTUNUS extension, ACR 2026 late-breaking abstract LB14
  26. [R51] argenx: UNITY discontinued after interim futility analysis (8 October 2026)
  27. [R52] Mariette et al. NECESSITY, ACR 2026 late-breaking abstract LB15
  28. [R56] ClinicalTrials.gov: NEPTUNUS extension, NCT05985915
  29. [R57] Amgen: OASIZ-301 results and ongoing OASIZ program (11 November 2026)
  30. [R58] Amgen trial record: OASIZ-301 enrollment field

Personal educational publication by Olga Goodman, MD. This is not a clinic portal or an emergency service. Any private expert service is separate from clinic accounts and insurance billing. Reuse requires attribution and a link to the original publication.

dr.rheumaview.com · Olga Goodman, MD

Share this page

FacebookLinkedInXWhatsAppTelegram

Topics: Clinical evidenceClinical trialsSjögren disease