TIPS in the era of ageing and metabolic comorbidity: a contemporary narrative review of risk, resilience, and outcomes
A contemporary narrative review identifies that age ≥70 years and metabolic comorbidities are consistently associated with unfavourable post-TIPS outcomes — higher 90-day mortality, higher hepatic encephalopathy incidence, and more readmissions — but the underlying evidence base is limited by retrospective designs and severe methodological heterogeneity.
| Population | Adults with advanced cirrhosis and portal hypertension, particularly older adults (≥70 years) with high prevalence of MASH, chronic kidney disease, sarcopenia, and frailty |
|---|---|
| Intervention | TIPS (transjugular intrahepatic portosystemic shunt), with or without systematic stent-graft under-dilation |
| Comparator | Younger adults undergoing TIPS or standard conservative management (null for direct comparator in some included studies) |
| Outcome | 90-day mortality (≥70 vs <70 years); 1-year mortality (≥70 vs <70 years); Post-TIPS overt hepatic encephalopathy; TIPS-associated hospital readmissions within 12 months; Transplant-free survival (≥70 vs <70 years); 30-day mortality (≥70 vs <70 years); Post-TIPS acute kidney injury |
Summary of findings
| Outcome | Effect | 95% CI | Certainty | Clinical relevance | Notes |
|---|---|---|---|---|---|
| 90-day mortality (≥70 vs <70 years) | direction unfavourable for age ≥70; in the pooled RR or OR with 95% CI reported across included studies | — | Low | — | 4 studies |
| 1-year mortality (≥70 vs <70 years) | direction unfavourable for age ≥70; in the pooled RR or OR with 95% CI reported | — | Low | — | 2 studies |
| Post-TIPS overt hepatic encephalopathy | direction unfavourable for age ≥70; in the pooled OR or HR with 95% CI reported uniformly | — | Low | — | 5 studies |
| TIPS-associated hospital readmissions within 12 months | direction unfavourable for age ≥70; in the pooled RR with 95% CI reported | — | Low | — | 1 studies |
| Transplant-free survival (≥70 vs <70 years) | no statistically significant difference up to 5 years in one externally validated cohort (derivation n=411, validation n=415); in the HR or 95% CI reported | — | Low | — | 1 studies |
| 30-day mortality (≥70 vs <70 years) | comparable between age groups at 30 days in most included studies; in the pooled RR or OR with 95% CI reported | — | Low | — | 3 studies |
| Post-TIPS acute kidney injury | discussed as mechanistically relevant in older patients; in the comparative quantitative data with 95% CI reported in included studies | — | Very low | — |
Context
The demographic profile of TIPS candidates has shifted toward older adults with MASH, chronic kidney disease, sarcopenia, and frailty. Current prognostic models (MELD, Child-Pugh) do not capture this multidimensional complexity. Randomised trials and prospective cohorts systematically under-represent patients ≥70 years, constraining evidence quality.
What the study showed
Retrospective studies included (combined n per cohort: 225–593 patients) consistently show that patients ≥70 years have higher 90-day and 1-year mortality, higher prevalence of MASH as aetiology, and more frequent refractory ascites as indication compared with younger patients. One study with 451 patients (50 propensity-score-matched per group) reported comparable 30-day in-hospital mortality but significantly higher mortality at later time points in older adults. An externally validated model (derivation n=411, validation n=415) incorporating creatinine and sodium predicted liver-related mortality after TIPS; systematic stent under-dilation was applied more frequently in ≥70-year-olds and may have attenuated acute haemodynamic adverse events. Absolute risk differences and 95% CIs are not consistently reported across primary studies.
How it was done
Contemporary narrative review based on primarily retrospective primary studies (5 retrospective cohorts, 1 prospectively collected with retrospective analysis; n per study: 136–593 patients; follow-up range: 30 days to >5 years). No systematic search with PRISMA criteria, no quantitative synthesis, and no formal risk-of-bias assessment (RoB 2/ROBINS-I) were performed. Covered stent-grafts were the dominant device in more recent studies.
Effect magnitude
Quantitative effect estimates with 95% CIs are not uniformly reported; the direction is consistently unfavourable for adults ≥70 years regarding medium-term mortality and hepatic encephalopathy, but no standardised effect sizes are available for synthesis.
Risk of bias
Absence of systematic search and formal risk-of-bias assessment (ROBINS-I not applied); all primary studies are retrospective with incomplete reporting of comorbidities, aetiology, TIPS indication, device characteristics, and haemodynamic response. Older adults are systematically under-represented in RCTs. Heterogeneity in age thresholds (65 vs 70 years), encephalopathy definitions, and selection criteria limits comparability. Selection bias ('carefully selected' older patients) may underestimate true procedural risk.
What this study does NOT prove
This narrative review does not prove causality between age and post-TIPS outcomes and does not establish age or comorbidity thresholds that contraindicate the procedure. Findings are not generalisable to non-Western populations, predominantly cholestatic cirrhosis, or settings with limited access to covered stent-grafts.
In clinical practice
Assessment of TIPS candidates aged ≥70 years requires multidimensional stratification including cardiovascular reserve, renal function (creatinine, sodium), frailty, and baseline cognitive status — parameters not captured by MELD alone. Stent-graft under-dilation may be considered in this group to attenuate acute haemodynamic stress, but lacks prospective evidence. Pre-TIPS hepatic encephalopathy (covert or overt) must be actively investigated, including formal cognitive screening before the procedure.
Limitations
Absence of systematic search and formal risk-of-bias assessment (ROBINS-I not applied); all primary studies are retrospective with incomplete reporting of comorbidities, aetiology, TIPS indication, device characteristics, and haemodynamic response. Older adults are systematically under-represented in RCTs. Heterogeneity in age thresholds (65 vs 70 years), encephalopathy definitions, and selection criteria limits comparability. Selection bias ('carefully selected' older patients) may underestimate true procedural risk.
What is still missing
Randomised controlled trials or dedicated prospective cohorts comparing TIPS versus standard of care in adults ≥70 years, with quality-of-life outcomes and systematic assessment of frailty, sarcopenia, and cardiac reserve.
Technical appendix
Version history
- 1.0 · 2026-11-01 — Auto-generated under Evidence Standard v1.0