Network meta-analysis

Serious adverse events

Metabolic dysfunction associated steatotic liver disease or steatohepatitis

longest follow-upRecorded harmsDoses of a drug pooledHazard ratio (from event counts)

Randomized trials

Treatments

866

Participants

Events

In plain language

Across 5 trials with 866 participants, 0 of 3 treatments had a clearly better result than placebo on serious adverse events, and for 3 the evidence could not tell them apart from placebo. The most favorable estimate was for Survodutide injection: hazard ratio (from event counts) 0.87 (95% credible interval 0.27 to 2.84).

Lower values favor the treatment. "Clearly" means the 95% credible interval excludes no difference. These are averages across trials; they do not say how any one person will respond. Not medical advice.

Network of trials

Each circle is a treatment; each line joins treatments that were randomized against each other in at least one trial.

Node area shows participants; line width shows trials comparing two treatments directly. Blue: GLP-1 receptor agonist containing treatments; grey: comparators. Hover over or tab to a node or line for details.

Effects compared with Placebo

Hazard ratio (from event counts). Lower values favor the treatment.

Versus Placebo, log scale. Blue: Bayesian random effects posterior median with 95% credible interval (thick) and 95% prediction interval (dashed). Amber diamond: frequentist estimate with 95% confidence interval. Hover over or tab to a row for its values.

Effects of each treatment versus Placebo
Treatment	Bayesian HR (95% CrI)	95% prediction interval	Frequentist HR (95% CI)	Verdict
Survodutide injection SURV_SC	0.87 (0.27 to 2.84)	0.19 to 3.84	0.85 (0.34 to 2.13)	Unclear
Tirzepatide injection TIRZ_SC	1.03 (0.25 to 4.96)	0.19 to 6.50	1.02 (0.26 to 3.91)	Unclear
Semaglutide injection SEMA_SC	1.57 (0.69 to 3.62)	0.48 to 5.17	1.64 (0.85 to 3.19)	Unclear

Ranking

How likely each treatment is to be among the best. Rankings ignore the size of differences and their certainty, so read them with the effects above.

League table

Every treatment against every other: hazard ratio (from event counts) of the row treatment versus the column treatment, posterior median and 95% credible interval. Green: row clearly better; red: row clearly worse.

Every treatment in a row compared with every treatment in a column
Row versus column	PLACEBO	SURV_SC	TIRZ_SC	SEMA_SC
PLACEBO	PLACEBO	1.14 0.35 to 3.75	0.98 0.20 to 3.96	0.64 0.28 to 1.44
SURV_SC	0.87 0.27 to 2.84	SURV_SC	0.84 0.11 to 5.56	0.55 0.13 to 2.47
TIRZ_SC	1.03 0.25 to 4.96	1.19 0.18 to 8.94	TIRZ_SC	0.66 0.13 to 3.80
SEMA_SC	1.57 0.69 to 3.62	1.81 0.40 to 7.92	1.51 0.26 to 7.94	SEMA_SC

Heterogeneity, fit and convergence

Bayesian random effects (primary)

Heterogeneity (tau): 0.298 (95% CrI 0.077 to 1.009), log scale
Residual deviance: 8.4 for 10 data points
DIC: 16.7 (pD 8.3)
Largest R-hat: 1.001
Smallest ESS: bulk 4,310, tail 4,134
Divergent transitions: 0
Convergence: All criteria met
Priors: PR02 N(0, 5^2); PR05 N(0, 2.5^2); PR09 log normal(-1.87, 1.52^2) on tau^2 (Adverse events, pharm vs placebo control)

Frequentist check (netmeta)

Method: inverse variance, REML, OR
Tau: 0.000
I squared: 0%
Q (total): 0.09 on 2 df, P = 0.956
Q (between designs): 0.00

Inconsistency

Do trials comparing two treatments directly agree with the indirect evidence through other treatments?

No comparison here has both direct and indirect evidence, so inconsistency cannot be checked by node splitting; the between designs Q above is the global check.

Small study effects

A comparison adjusted funnel plot: each point is one trial comparison, centered on its own network estimate. Asymmetry can signal that small trials with unfavorable results are missing.

Fewer than ten trial comparisons, too few to judge funnel plot asymmetry.

Comparability of the trials

Checked across designs (the set of treatments each trial compares) for eligibility, outcome definition and assessment, and the baseline factors above. Flags point to possible intransitivity; notes describe variation that is not tied to design.

NoteOutcome assessment, time point: selected time points span 52 to 79 weeks
NoteOutcome assessment, analysis set: mixed: randomized 1, safety 4
NoteData handling, input conversions: arms_combined 2

Comparability items per trial
Trial	time weeks	analysis set	definition	instrument	adjudicated	proportion method	intercurrent events (discontinuation)	intercurrent events (rescue)	blinding	funding	start year	Eligibility
SYNERGY-NASH	56	randomized	not stated	not stated	not stated	not stated	not stated	not stated	double_blind	industry	2019	age 18 to 80, BMI 27 to 50, biopsy MASH with NAS 4 or more and fibrosis F2 to F3, HbA1c up to 9.5 percent
NCT03357380	79	safety	not stated	not stated	not stated	not stated	not stated	not stated	double_blind	industry	2017	age 18 to 75; BMI 25 to 40; MRE stiffness 2.50 to 4.63 kPa; MRI PDFF 10 percent or more; HbA1c 9.5 or less
NCT02970942	79	safety	not stated	not stated	not stated	not stated	not stated	not stated	double_blind	industry	2016	age 18 to 75; biopsy NASH F1 to F3; NAS 4 or more; BMI above 25; HbA1c 10 percent or less
NCT03987451	55	safety	not stated	not stated	not stated	not stated	not stated	not stated	double_blind	industry	2019	age 18 to 75; biopsy NASH with F4 compensated cirrhosis; NAS 3 or more; BMI 27 or more
NCT06309992	52	safety	not stated	not stated	not stated	not stated	not stated	not stated	double_blind	industry	2024	age 18 or more; BMI 30 or more, or 27 with complication; MRI-PDFF 8 percent or more plus NIT or biopsy MASH

Trial data

What each trial contributes. Arm rows give the arm value with its standard error, or events over participants; contrast rows give the hazard ratio (from event counts) on the analysis scale (log) against the trial's first row. Flags mark converted, combined or approximated inputs.

Trial data entering this analysis
Trial	Arm	Treatment	Week	Data	Events / N	Flags	Source
SYNERGY-NASH	A1+A2+A3	TIRZ_SC	56	arm, events	9 / 142	arms_combined	R0018381:Table 3
SYNERGY-NASH	A4	PLACEBO	56	arm, events	3 / 48		R0018381:Table 3
NCT03357380	A1	SEMA_SC	79	arm, events	4 / 34		R0011165:Table 3
NCT03357380	A2	PLACEBO	79	arm, events	3 / 33		R0011165:Table 3
NCT02970942	A1+A2+A3	SEMA_SC	79	arm, events	39 / 240	arms_combined	registry:adverse_events
NCT02970942	A4	PLACEBO	79	arm, events	8 / 80		registry:adverse_events
NCT03987451	A1	SEMA_SC	55	arm, events	6 / 47		R0017043:Table 2
NCT03987451	A2	PLACEBO	55	arm, events	2 / 24		R0017043:Table 2
NCT06309992	A1	SURV_SC	52	arm, events	14 / 146		R0011245:Table 3
NCT06309992	A2	PLACEBO	52	arm, events	8 / 72		R0011245:Table 3

Rerun this analysis yourself

Remove treatments or trials and refit the frequentist and Bayesian models in your browser, with the same methods as the published analysis.

Open in the NMA lab