← Back

SMART-MS

Intrathecal Mesenchymal Stem Cells in Progressive Multiple Sclerosis: A Randomized, Double-Blind, Placebo-Controlled Trial (SMART-MS)

Share Share Copied! Compare

Year of Publication: 2026

Authors: Kvistad CE, Kråkenes T, Holmøy T, ..., Bø L

Journal: Neurology

Citation: Neurology 2026;106:e214915. doi:10.1212/WNL.0000000000214915

Link: https://doi.org/10.1212/WNL.0000000000214915

Clinical Question

Does a single intrathecal injection of autologous bone marrow-derived mesenchymal stem cells produce neuroregenerative effects in progressive multiple sclerosis?

Bottom Line

A single intrathecal injection of autologous MSCs did NOT produce a measurable neuroregenerative effect in progressive MS, and was associated with concerning local inflammatory adverse events including arachnoiditis. Intrathecal MSC administration in progressive MS should be approached with caution in future studies.

        Major Points
        Negative primary outcome: no significant between-group difference in change of combined evoked potential latency at 6 months (β = −0.31, 95% CI −1.84 to 1.22, p = 0.668)
Reduced cerebral atrophy on MRI at 6 months in MSC group (β = 9.37, 95% CI 0.29 to 18.45, p = 0.044), not sustained at 12 months
Lower serum GFAP at 6 months (β = −16.3 pg/mL, 95% CI −33.0 to 0.3, p = 0.054), not sustained at 12 months
Exploratory CSF proteomics showed reductions in multiple inflammation-related proteins at 6 months
Notable safety signal: one SAE probably related to MSC, fever (n=9), low back pain (n=10), spinal MRI abnormalities with fluid loculations and nerve root clumping (n=7), one case of chronic coccygeal pain attributed to arachnoiditis
Class III evidence that intrathecal autologous MSCs do NOT provide neuroregenerative effects in progressive MS

      

Design

Study Type: Randomized, double-blind, placebo-controlled, crossover phase I/II trial

Randomization: 1

Blinding: Double-blind; study drug prepared in separate room and injected behind patient; all patients underwent bone marrow aspiration twice to maintain blinding

Allocation: Concealed; blocks of 4 using computer-generated sequence prepared by an independent statistician

Enrollment Period: First patient enrolled August 9, 2021

Follow-up Duration: 18 months (end-of-study safety visit); efficacy outcomes at 6 and 12 months

Centers: 4

Countries: Norway

Sample Size: 18

Analyzed: 18

Analysis: Baseline-adjusted regression (ANCOVA) models; missing evoked potential data imputed using longest conduction time within same modality

Power Calculation: Sample size of 18 patients determined based on feasibility and precedent from similar early-phase proof-of-concept trials

Registration: ClinicalTrials.gov NCT04749667 (registered February 8, 2021)

Inclusion Criteria

Primary or secondary progressive MS per revised McDonald criteria
Age 18-55 years
Expanded Disability Status Scale (EDSS) score of 4-7
Disease duration of 2-18 years
No relapses for at least 24 months before enrollment
No new or enlarging MRI lesions for at least 24 months before enrollment
Untreated with DMTs during the study

Exclusion Criteria

Active or chronic infection
History of malignancy
Polyneuropathy
Severe comorbid illness
Other conditions that could affect cognition or compliance

Arms

Field	Arm A: MSC then Placebo	Control
N	0	0
Intervention	Autologous intrathecal MSCs 1 × 10^6 cells/kg (max 100 × 10^6) at baseline, then 0.9% saline placebo at 6 months (crossover)	0.9% saline placebo at baseline, then autologous intrathecal MSCs 1 × 10^6 cells/kg at 6 months (crossover)
Duration	18 months total follow-up	18 months total follow-up

Outcomes

Outcome	Type	Control	Intervention	P-value
Change in latency of combined evoked potentials (CEPs) — sum of z-transformed latencies for visual (VEP), motor (MEP), and somatosensory (SEP) evoked potentials bilaterally	Primary	Placebo group	MSC group	0.668
	Secondary			0.044
	Secondary			0.054
	Secondary
	Secondary
	Safety
	Safety
	Safety
	Safety
	Safety
Fever (post-MSC)	Adverse
Low back pain (post-MSC)	Adverse
Spinal MRI abnormalities (fluid loculations, nerve root clumping)	Adverse
Arachnoiditis with chronic coccygeal pain	Adverse
Serious AE probably MSC-related	Adverse

Criticisms

Small sample size (n=18) limits interpretation and statistical power, as acknowledged by authors
Crossover design may complicate interpretation of 12-month outcomes given that placebo arm received MSCs at 6 months
Single-country (Norway) enrollment may limit generalizability
Notable safety signals (arachnoiditis, fluid loculations, nerve root clumping in 7/18) raise concerns about intrathecal MSC safety profile
Class III level of evidence

Funding

Article Processing Charge funded by Klinbeforsk

Based on: SMART-MS (Neurology, 2026)

Authors: Kvistad CE, Kråkenes T, Holmøy T, ..., Bø L

Citation: Neurology 2026;106:e214915. doi:10.1212/WNL.0000000000214915

Content summarized and formatted by NeuroTrials.ai.

SMART-MS

(2026)

Objective

To assess whether a single intrathecal administration of autologous bone marrow-derived mesenchymal stem cells (MSCs) could provide evidence of a neuroregenerative effect in patients with progressive multiple sclerosis.

Study Summary

• No significant between-group difference for the primary end point of change in combined evoked potential latency at 6 months (β = −0.31, 95% CI −1.84 to 1.22, p = 0.668)
• MSC group showed reduced cerebral atrophy on MRI at 6 months (β = 9.37, 95% CI 0.29 to 18.45, p = 0.044) and lower serum GFAP (β = −16.3 pg/mL, 95% CI −33.0 to 0.3, p = 0.054), but neither sustained at 12 months
• Adverse events suggested an acute localized inflammatory reaction: fever (n=9), low back pain (n=10), spinal MRI abnormalities with fluid loculations and nerve root clumping (n=7), and one case of chronic coccygeal pain from arachnoiditis
• Class III evidence that a single intrathecal autologous MSC dose does NOT provide a neuroregenerative effect in progressive MS

Intervention

Single intrathecal injection of autologous bone marrow-derived mesenchymal stem cells (1 × 10^6 cells/kg, max 100 × 10^6) vs 0.9% saline placebo, with crossover at 6 months.

Inclusion Criteria

Adults aged 18-55 with primary or secondary progressive MS by revised McDonald criteria, EDSS 4-7, disease duration 2-18 years, no relapses and no new/enlarging MRI lesions for ≥24 months, untreated with DMTs during the study.

Study Design

Arms: Autologous intrathecal MSCs 1×10^6/kg then placebo crossover at 6 months vs Placebo (saline) then MSCs crossover at 6 months (total N=18)

Patients per Arm: 18 total patients in crossover design

Outcome

• Primary end point negative: change in combined evoked potential latency at 6 months β = −0.31 (95% CI −1.84 to 1.22, p = 0.668)
• Secondary: reduced cerebral atrophy at 6 months (β = 9.37, 95% CI 0.29 to 18.45, p = 0.044), lower serum GFAP (β = −16.3 pg/mL, p = 0.054); not sustained at 12 months
• Exploratory CSF proteomics: reductions in multiple inflammation-related proteins at 6 months
• Safety concerns: one SAE probably related to MSC, common AEs of fever and low back pain, spinal MRI abnormalities in 7 patients, one case of arachnoiditis with chronic coccygeal pain

Bottom Line

Major Points

Negative primary outcome: no significant between-group difference in change of combined evoked potential latency at 6 months (β = −0.31, 95% CI −1.84 to 1.22, p = 0.668)
Reduced cerebral atrophy on MRI at 6 months in MSC group (β = 9.37, 95% CI 0.29 to 18.45, p = 0.044), not sustained at 12 months
Lower serum GFAP at 6 months (β = −16.3 pg/mL, 95% CI −33.0 to 0.3, p = 0.054), not sustained at 12 months
Exploratory CSF proteomics showed reductions in multiple inflammation-related proteins at 6 months
Notable safety signal: one SAE probably related to MSC, fever (n=9), low back pain (n=10), spinal MRI abnormalities with fluid loculations and nerve root clumping (n=7), one case of chronic coccygeal pain attributed to arachnoiditis
Class III evidence that intrathecal autologous MSCs do NOT provide neuroregenerative effects in progressive MS

Study Design

Study Type: Randomized, double-blind, placebo-controlled, crossover phase I/II trial
Randomization: Yes
Blinding: Double-blind; study drug prepared in separate room and injected behind patient; all patients underwent bone marrow aspiration twice to maintain blinding
Sample Size: 18
Follow-up: 18 months (end-of-study safety visit); efficacy outcomes at 6 and 12 months
Centers: 4
Countries: Norway

Primary Outcome

Definition: Change in latency of combined evoked potentials (CEPs) — sum of z-transformed latencies for visual (VEP), motor (MEP), and somatosensory (SEP) evoked potentials bilaterally

Control	Intervention	HR/OR	P-value
Placebo group	MSC group	- (−1.84 to 1.22)	0.668

Limitations & Criticisms

Small sample size (n=18) limits interpretation and statistical power, as acknowledged by authors
Crossover design may complicate interpretation of 12-month outcomes given that placebo arm received MSCs at 6 months
Single-country (Norway) enrollment may limit generalizability
Notable safety signals (arachnoiditis, fluid loculations, nerve root clumping in 7/18) raise concerns about intrathecal MSC safety profile
Class III level of evidence

Citation

Neurology 2026;106:e214915. doi:10.1212/WNL.0000000000214915

View Original Publication →

SMART-MS

Intrathecal Mesenchymal Stem Cells in Progressive Multiple Sclerosis: A Randomized, Double-Blind, Placebo-Controlled Trial (SMART-MS)

Clinical Question

Bottom Line

Major Points

Design

Inclusion Criteria

Exclusion Criteria

Arms

Outcomes

Criticisms

Funding

Ask about SMART-MS