GRADE Summary of Findings.
GRADE Summary of Findings.
| Outcome | Effect/Direction | Evidence stream | Certainty (GRADE) | Reason for Downgrading |
|---|---|---|---|---|
| Suicide death and self-harm (adults with T2DM / obesity) | No class-wide increase; rare events; near-null pooled estimates | RCT meta-analyses; active-comparator cohorts; regulatory reviews | Moderate | Imprecision (rare events) |
| Incident depression | No class-wide increase; some comparator-sensitive variation | RCT meta-analyses; active-comparator cohorts | Low | Indirectness (heterogeneous outcome definitions); imprecision |
| Anxiety symptoms / disorder | No class-wide increase | RCT meta-analyses; active-comparator cohorts | Low | Indirectness (heterogeneous ascertainment); imprecision (few dedicated analyses) |
| Tirzepatide-specific psychiatric outcomes | No signal of increased risk | Post hoc pooled RCT (SURMOUNT); one real-world cohort | Low | Imprecision (few studies); indirectness (limited follow-up) |
| Pharmacovigilance disproportionality signals | Mixed signals; stronger signals in subgroups with co-reported antidepressants or benzodiazepines | EudraVigilance; FAERS; WHO VigiBase | Very low | No denominator data; reporting and notoriety bias; cannot estimate incidence |
GRADE certainty ratings: high, moderate, low, very low. Downgrading domains: risk of bias, inconsistency, indirectness, imprecision, publication bias. Regulatory reviews are not given a separate row in this table because they are not assigned GRADE certainty ratings; their conclusions integrate multiple evidence streams under a formal risk–benefit framework and are summarized in Table 2 and in the Regulatory Evidence section.
EudraVigilance, European Union adverse event reporting database; FAERS, FDA Adverse Event Reporting System; GRADE, Grading of Recommendations Assessment, Development and Evaluation; RCT, randomized controlled trial; T2DM, type 2 diabetes mellitus; VigiBase, World Health Organization global pharmacovigilance database; WHO, World Health Organization.