Publication

• Title: Total Intravenous vs Volatile Inhalational Anesthesia for Major Noncardiac Surgery: A Randomized Clinical Trial
• Acronym: VITAL
• Year: 2026
• Journal published in: JAMA
• Citation: Jhanji S, Booth K, Hiller L, et al; VITAL Trial Team. Total intravenous vs volatile inhalational anesthesia for major noncardiac surgery: a randomized clinical trial. JAMA. Published online August 12, 2026.

Context & Rationale

• Background

  • General anaesthesia for major non-cardiac surgery is commonly maintained with either a volatile inhalational agent or a propofol-based total intravenous anaesthesia technique.
  • The techniques have different pharmacological, inflammatory, haemodynamic and environmental profiles. Volatile agents have been associated with putative cardioprotective effects, whereas propofol has been associated with reduced postoperative nausea and vomiting, faster early recovery and possible anti-inflammatory effects.
  • Observational studies had generated competing hypotheses about anaesthetic technique and long-term cancer outcomes, while concerns about greenhouse-gas emissions had accelerated the move away from volatile agents in some health systems.
  • A 2024 meta-analysis of 317 randomised trials involving more than 51,000 participants found no convincing difference in mortality or major organ morbidity, but favoured TIVA for several early patient-centred recovery outcomes. The underlying trials were heterogeneous and did not replace the need for a large pragmatic comparison in the broad non-cardiac surgical population.¹
• Research Question/Hypothesis

  • Among patients aged at least 50 years undergoing elective major non-cardiac surgery, would maintenance of general anaesthesia with propofol-based TIVA increase days alive and at home during the first 30 postoperative days compared with volatile inhalational anaesthesia?
  • The superiority hypothesis was that TIVA would improve patient-centred recovery, survival and perioperative safety.
• Why This Matters

  • Both techniques are used in millions of patients, so even a modest difference in recovery, complications or survival could have substantial consequences for patients and health services.
  • Use of TIVA had expanded faster than the generation of large-scale safety data, particularly in older adults undergoing non-cardiac surgery.
  • A clinically similar outcome profile would allow anaesthetic choice to be guided more confidently by individual patient priorities, postoperative symptom risk, clinician expertise, equipment, cost and environmental considerations.

Design & Methods

• Research Question: Does propofol-based TIVA improve days alive and at home at 30 days compared with volatile inhalational anaesthesia in adults aged at least 50 years undergoing elective major non-cardiac surgery?
• Study Type: Pragmatic, multicentre, parallel-group, open-label, superiority randomised clinical trial embedded within the UK Perioperative Quality Improvement Programme and conducted at 49 National Health Service hospitals.
• Population:
  • Adults aged at least 50 years.
  • Scheduled for elective major non-cardiac surgery under general anaesthesia.
  • Written informed consent obtained before randomisation.
  • Excluded for a known contraindication to either technique, clinician refusal, inability to provide consent or complete questionnaires, previous VITAL participation, or an expectation that the patient would not survive for 30 days after surgery.
  • Randomisation occurred on the day of surgery using a computerised minimisation algorithm balancing surgical specialty, expected procedure duration of less than 2 hours versus at least 2 hours, cancer versus non-cancer surgery, and preoperative Rockwood Clinical Frailty Scale category.
• Intervention:
  • Maintenance of general anaesthesia using an intravenous propofol infusion as the intended primary maintenance technique.
  • No minimum propofol dose, target concentration or depth-of-anaesthesia target was mandated.
  • Induction drugs, opioids, neuromuscular blockade, airway management, regional analgesia, antiemetics, monitoring and all other perioperative care remained at the treating anaesthetist’s discretion.
• Comparison:
  • Maintenance of general anaesthesia using a volatile inhalational agent, principally sevoflurane or isoflurane, according to local practice.
  • No minimum end-tidal concentration or anaesthetic depth target was specified.
  • All other aspects of perioperative management were delivered at clinician discretion, as in the TIVA group.
• Blinding: Treating clinicians and site research teams were not blinded. Participants were not explicitly informed of allocation, but were not formally blinded. The objective primary outcome was relatively resistant to ascertainment bias, whereas postoperative symptom and satisfaction outcomes were more vulnerable to performance and reporting bias.
• Statistics: A total of 2,500 participants was required to detect a 1-day difference in days alive and at home at 30 days, assuming a mean of 25 days and standard deviation of 7.5 days, with 90% power at a two-sided 5% significance level and allowance for 5% loss to follow-up. The primary available-case intention-to-treat analysis used zero-inflated negative-binomial regression adjusted for the four minimisation variables; a per-protocol analysis was prespecified. Secondary analyses were not adjusted for multiplicity.²
• Follow-Up Period: Outcomes were assessed from postoperative day 1 to 6 months, with the primary outcome at 30 days and key secondary assessments at day 1, day 3, day 30, day 90 and 6 months.

Key Results

This trial was not stopped early. It completed planned recruitment, randomising 2,508 participants equally between the two groups.

• The primary estimate was essentially zero, and its confidence interval excluded both the prespecified 1-day superiority target and the stated 3-day minimum clinically important difference.
• The per-protocol analysis was concordant with the primary analysis: adjusted IRR 1.00; 95% CI 0.99 to 1.02; P=0.54. No convincing treatment interaction was apparent across surgical specialty, procedure duration, cancer status, frailty or the participating hospital’s previous use of TIVA.
• The day-1 symptom differences may be clinically relevant to individual patients, particularly those at high risk of postoperative nausea and vomiting, but they were exploratory and did not translate into better global recovery, fewer complications or more days at home.

Internal Validity

• Randomisation and Allocation: Central computerised randomisation on the day of surgery used minimisation to balance four important prognostic variables. Allocation was concealed until the participant was entered into the randomisation system, and baseline characteristics were closely balanced.
• Dropout and Post-randomisation Exclusions: One participant withdrew consent for all data. The primary outcome was available for 1,225 of 1,253 participants assigned to TIVA and 1,221 of 1,254 assigned to inhalational anaesthesia. The 61 participants omitted from the primary analysis comprised 40 withdrawals before day 30, 15 with missing primary-outcome data and 6 with incorrect readmission data. The small and balanced loss reduces, but does not eliminate, the departure from a complete intention-to-treat analysis.
• Performance and Detection Bias: Clinician blinding was impossible. The primary outcome depended principally on mortality, discharge and readmission data and was therefore comparatively robust. Day-1 discomfort, satisfaction and quality-of-recovery outcomes were more susceptible to differences in care, expectations and reporting.
• Protocol Adherence and Crossover: Adherence was high and balanced: 1,205/1,253 participants assigned to TIVA (96.2%) and 1,197/1,254 assigned to inhalational anaesthesia (95.5%) received their allocated maintenance strategy. Eight participants received both techniques, 60 received only the alternative technique, one received no general anaesthetic and treatment status was unclear in 36.
• Baseline Characteristics: Groups were comparable in age, sex, frailty, surgical specialty, cancer status and expected procedure duration. However, the overall population was relatively low risk: 80.5% were classified as well and only 4% as frail, contributing to lower mortality than anticipated.
• Heterogeneity: Inclusion of several surgical specialties strengthened the pragmatic question, and minimisation reduced prognostic imbalance. Nevertheless, 72.1% underwent intra-abdominal surgery and 75.4% underwent cancer surgery, so the numerical dominance of these groups limited the information contributed by vascular, musculoskeletal and thoracic subgroups.
• Timing: Participants were randomised on the day of surgery and the allocated strategy was applied during maintenance of the index anaesthetic. This was appropriate for a brief intraoperative intervention intended to influence subsequent recovery.
• Dose: No fixed propofol infusion rate, target-controlled concentration, minimum alveolar concentration or anaesthetic-depth target was mandated. This improved pragmatism but means the trial evaluated two clinical maintenance strategies rather than standardised pharmacological exposures.
• Separation of the Variable of Interest: Propofol infusion was used in 96.3% of the TIVA group versus 3.0% of the inhalational group. A volatile maintenance agent was used in 2.2% versus 95.7%, respectively. This represents strong separation of the intended maintenance technique.
• Key Delivery Aspects: Important parts of the delivered anaesthetic bundle differed: opioid infusion was used in 83.6% of the TIVA group versus 44.8% of the inhalational group, and depth-of-anaesthesia monitoring in 90.6% versus 49.5%. These differences reflect real-world implementation but complicate attribution of early symptom differences solely to propofol or the volatile agent.
• Adjunctive Therapy Use: Other major co-interventions were well balanced, including regional analgesia (78.2% versus 78.3%), intravenous dexamethasone (77.8% in both groups), muscle relaxant use (93.5% versus 94.8%) and reversal of neuromuscular blockade (65.4% versus 65.6%).
• Outcome Assessment: DAH30 was clearly defined and patient-centred, integrating death, acute hospital stay and readmission. Its interpretation is not identical to functional recovery, because any location outside an acute hospital counted as home. Delirium was assessed only on postoperative day 3 and could therefore miss earlier, fluctuating or later episodes.
• Statistical Rigor: The primary model accommodated the markedly skewed and zero-inflated distribution and adjusted for minimisation variables. Raw descriptive results and the per-protocol analysis were concordant. Secondary outcomes were numerous and not adjusted for multiplicity, and the unexpectedly low mortality rate prevented completion of the intended 90-day mortality non-inferiority assessment.

Conclusion on Internal Validity: Internal validity is strong for the conclusion that TIVA does not materially improve average DAH30 in this population, supported by concealed randomisation, good treatment separation, low missingness and a very precise estimate. Validity is more moderate for subjective early-recovery outcomes and limited for rare safety outcomes such as mortality and unintentional awareness.

External Validity

• Population Representativeness: The mean age was 67 years, 55% were male, 75.4% underwent cancer surgery, 72.1% underwent intra-abdominal surgery and 98.1% had procedures expected to last at least 2 hours. Most participants were physiologically well; only 4% were classified as frail. White British participants comprised 86% to 88% of each group.
• Recruitment Selection: Of 6,143 eligible patients, 2,508 were randomised. Reasons for non-recruitment included participant refusal, staff unavailability, clinician refusal and unrecorded or other reasons. These practical exclusions may have selected patients and surgical pathways more amenable to research participation.
• Applicability to Contemporary Practice: The pragmatic intervention, high protocol adherence and inclusion of hospitals with widely varying pretrial TIVA use support application across comparable UK elective surgical services.
• Healthcare-System Effects: Regional anaesthesia was used in approximately 78% of both groups, and care was delivered within the UK NHS and PQIP infrastructure. Length of stay, rehabilitation pathways, availability of processed electroencephalography and anaesthetic staffing may differ elsewhere.
• Populations Not Addressed: The findings should not be directly extrapolated to emergency surgery, cardiac surgery, minor or ambulatory procedures, younger patients, severely frail or cognitively impaired patients, or settings with substantially different perioperative pathways.
• Oncological Applicability: Although three-quarters underwent cancer surgery, follow-up was limited to 6 months and the trial did not assess recurrence, disease-free survival or longer-term cancer-specific survival.

Conclusion on External Validity: External validity is good for patients resembling the predominantly fit, older UK population undergoing prolonged elective intra-abdominal or cancer surgery. Generalisability is appreciably narrower for emergency surgery, very frail patients, other healthcare systems and less well-represented surgical specialties.

Strengths & Limitations

• Strengths:
  • Large, pragmatic comparison conducted across 49 NHS hospitals.
  • Central concealed randomisation with minimisation across clinically important prognostic variables.
  • Excellent separation of the allocated maintenance techniques and overall adherence of 95.8%.
  • Patient-selected, objective and clinically interpretable primary outcome incorporating death, hospitalisation and readmission.
  • Narrow primary confidence interval that excludes even a small average benefit from TIVA.
  • Concordant intention-to-treat and per-protocol analyses.
  • Detailed collection of intraoperative practice, complications, postoperative symptoms, delirium and awareness.
  • Embedding within PQIP reduced research burden and supported efficient multicentre recruitment and follow-up.
  • Independent trial steering and data monitoring committees, public funding and limited declared conflicts of interest.
• Limitations:
  • Open-label delivery, with potential performance and reporting bias in subjective outcomes.
  • The primary analysis omitted participants without calculable DAH30 rather than retaining every randomised participant.
  • Low mortality and extremely rare awareness events prevented firm comparative safety conclusions.
  • Multiple secondary and symptom comparisons were not adjusted for multiplicity.
  • The intervention was not dose-standardised and differed from the comparator in opioid-infusion and depth-monitoring practices.
  • DAH30 counted any location outside an acute hospital as home and did not directly measure function or quality of life while at home.
  • Delirium was assessed at one postoperative time point rather than serially throughout the period of greatest risk.
  • The cohort was predominantly well, White British and undergoing intra-abdominal cancer surgery, with few severely frail patients.
  • Emergency surgery, long-term cancer outcomes, cost-effectiveness and environmental outcomes were not evaluated in the current report.

Interpretation & Why It Matters

• Primary Clinical Conclusion

  TIVA should not be selected routinely in the expectation that it will increase days alive and at home after major elective non-cardiac surgery. The primary result was both statistically and clinically precise.
• Immediate Patient Comfort

  TIVA was associated with numerically less nausea or vomiting, thirst and hoarseness on the first postoperative day. These differences may matter to individual patients but were exploratory and did not improve global quality of recovery or subsequent time at home.
• Major Clinical Outcomes

  Complications and mortality were numerically similar, with no consistent signal favouring either technique. The low event rates mean that absence of statistical significance should not be interpreted as definitive proof that the techniques have identical rare-event safety profiles.
• Choice of Anaesthetic Technique

  For similar elective surgical patients, either strategy is reasonable. Selection can be individualised according to previous postoperative nausea and vomiting, patient preference, clinician expertise, intravenous access, equipment, processed electroencephalography, theatre workflow and environmental priorities.
• What VITAL Changes

  The trial replaces mechanistic and observational arguments with a precise pragmatic estimate for overall 30-day recovery. It weakens claims that either maintenance technique confers a large general advantage in survival, complications or recovery after elective major non-cardiac surgery.

Controversies & Other Evidence

• Superiority is not the same as equivalence: VITAL was designed to show superiority of TIVA, not to prove that the two techniques are equivalent. For DAH30, however, the confidence interval was sufficiently narrow to exclude both the prespecified 1-day benefit and the stated 3-day minimum clinically important difference. That precision does not extend automatically to mortality, awareness or individual complications.
• The planned mortality non-inferiority analysis was inconclusive: The power calculation assumed 90-day mortality of 3.8%, but observed mortality was only 1.35%. The supplementary analysis estimated that approximately 7,232 participants would have been needed to exclude a relative risk of 1.5 using a one-sided 95% confidence interval, or 6,376 to exclude a relative risk of 1.6 using a one-sided 97.5% confidence interval. VITAL therefore provides reassuring observed mortality data but does not establish formal mortality non-inferiority.³
• Timing of the statistical analysis plan: The final statistical analysis plan was dated 11 December 2024, after final participant follow-up in October 2024. The supplied documents do not establish whether treatment-group outcome data had been examined or when the database was locked. The reported analyses largely accord with the plan, but public finalisation before completion of follow-up would have provided stronger protection against analytic flexibility.⁴
• Sample-size paragraph labelling error: The paper describes pilot data as having a mean “DAH30” of 72.9 days, which is impossible for an outcome bounded between 0 and 30. The protocol and statistical analysis plan identify 72.9 days as the mean DAH90 value. This is an evident labelling error rather than a substantive change to the power calculation.²⁴
• One-day target versus 3-day minimum clinically important difference: The trial was powered to detect a 1-day DAH30 difference, whereas the results table identifies 3 days as the minimum clinically important difference. The rationale for powering the trial below that threshold could have been clearer. Nonetheless, the observed 95% confidence interval of −0.52 to 0.54 days excludes either threshold and makes the primary inference secure.
• The randomised contrast was an anaesthetic strategy rather than a single drug: Strong separation of propofol and volatile maintenance was accompanied by much greater use of opioid infusion and processed depth monitoring with TIVA. These co-interventions are integral to real-world TIVA delivery, but they prevent a strictly pharmacological interpretation of the early symptom findings.
• Complex modelling of DAH30: The zero-inflated negative-binomial model was reasonable for a skewed outcome with excess zeros, but the inflation component was numerically unstable because death deterministically assigned a value of zero; the coefficient for death was 25.0 with a standard error exceeding 12,000. Confidence in the main conclusion instead comes from convergence of the adjusted count model, raw means and medians, the comparison of zero versus non-zero outcomes, and the per-protocol analysis.³
• Interpretation of days at home: DAH30 is meaningful and efficient, but “home” included care homes, rehabilitation facilities and any other location outside an acute hospital. In this low-mortality cohort, the outcome was driven predominantly by postoperative length of stay and readmission rather than survival or demonstrated functional independence.
• Early symptom findings remain exploratory: Adjusted treatment effects, confidence intervals and P values were not provided for the individual Bauer discomfort items, and no multiplicity adjustment was planned. The direction of effect is nevertheless consistent with the broader randomised literature showing less postoperative nausea and vomiting and some early recovery advantages with propofol-based TIVA.¹
• The propofol mortality debate: A 2023 meta-analysis reported a small increase in mortality with propofol across heterogeneous perioperative and critical care trials: RR 1.10; 95% CI 1.01 to 1.20.⁵ Its interpretation was challenged because of heterogeneous mortality time points, influential data extraction decisions and exclusion of the large MYRIAD trial.⁶ VITAL provides no signal of a large mortality disadvantage from TIVA, but its low event rate cannot resolve a relative effect as small as 10%.
• Consistency with other large trials: Large randomised comparisons in cardiac surgery and cancer surgery have likewise not demonstrated a consistent advantage of volatile or intravenous maintenance for mortality, major postoperative complications or long-term cancer survival.⁷⁸⁹¹⁰ VITAL extends this pattern to a broad elective non-cardiac surgical population using a patient-centred recovery outcome.
• Delirium evidence is not fully settled: In the SPOP1 randomised trial of older adults undergoing major cancer surgery, postoperative delirium occurred in 8.4% with propofol and 12.4% with sevoflurane: RR 0.68; 95% CI 0.48 to 0.95.¹¹ VITAL found similar day-3 delirium rates of 6.4% and 6.7%, but a single day-3 assessment is not directly comparable with serial surveillance. A 2026 global observational analysis associated propofol with more delirium, cognitive decline and mortality,¹² although its accompanying critique emphasised residual confounding and the greater evidential weight of randomised comparisons.¹³
• Environmental and economic conclusions require separate evidence: VITAL did not measure greenhouse-gas emissions, energy use, plastic waste, propofol disposal or lifecycle environmental impact. A health-economic evaluation was planned, but cost-effectiveness and longitudinal EQ-5D findings were not included in the current report.² The absence of a material DAH30 difference permits environmental and economic factors to enter decision-making; it does not demonstrate that TIVA is environmentally or economically superior.

Summary

• VITAL randomised 2,508 adults aged at least 50 years undergoing elective major non-cardiac surgery at 49 UK NHS hospitals to propofol-based TIVA or volatile inhalational maintenance.
• TIVA did not improve the primary outcome: mean DAH30 was 22.5 versus 22.4 days; adjusted IRR 1.00; 95% CI 0.99 to 1.02; P=0.68.
• DAH90, quality of recovery, major complications, delirium and mortality through 6 months were similar, although rare safety outcomes were inadequately powered for formal equivalence or non-inferiority conclusions.
• TIVA was associated with less day-1 nausea or vomiting, thirst and hoarseness, but these exploratory symptom differences did not improve broader or longer-term recovery.
• The primary conclusion has strong internal validity and applies well to comparable elective UK surgical patients, but not automatically to emergency surgery, very frail patients, other healthcare systems or long-term oncological outcomes.

Overall Takeaway

VITAL is a large, practice-informing comparative-effectiveness trial showing that propofol-based TIVA does not produce a clinically meaningful improvement in days alive and at home after elective major non-cardiac surgery compared with volatile inhalational anaesthesia. Its principal contribution is to remove anticipated differences in global recovery as a compelling reason to choose one technique universally, while leaving patient comfort, clinician expertise, operational factors and environmental priorities to guide individualised practice.

Bibliography

• 1.Kampman JM, Hermanides J, Hollmann MW, et al. Mortality and morbidity after total intravenous anaesthesia versus inhalational anaesthesia: a systematic review and meta-analysis. EClinicalMedicine. 2024;72:102636.
• 2.Yeung J, Jhanji S, Braun J, et al; VITAL Trial Team. Volatile vs Total Intravenous Anaesthesia for Major Non-Cardiac Surgery: a pragmatic randomised trial (VITAL). Trials. 2024;25:414.
• 3.Jhanji S, Booth K, Hiller L, et al; VITAL Trial Team. Supplemental online content for: Total intravenous vs volatile inhalational anesthesia for major noncardiac surgery: a randomized clinical trial. JAMA. Published online August 12, 2026.
• 4.VITAL Trial Team. Statistical analysis plan: Volatile vs Total Intravenous Anaesthesia for Major Non-Cardiac Surgery. Version 1.0. December 11, 2024.
• 5.Kotani Y, Pruna A, Turi S, et al. Propofol and survival: an updated meta-analysis of randomized clinical trials. Crit Care. 2023;27:139.
• 6.Glass A, McCall P, Shelley B. Comment on Propofol and survival: an updated meta-analysis of randomized clinical trials. Crit Care. 2023;27:281.
• 7.Landoni G, Lomivorotov VV, Nigro Neto C, et al; MYRIAD Study Group. Volatile anesthetics versus total intravenous anesthesia for cardiac surgery. N Engl J Med. 2019;380:1214–1225.
• 8.Deng XQ, Yu H, Wang WJ, et al. Effect of volatile vs propofol anaesthesia on major complications and mortality after cardiac surgery: a multicentre randomised trial. Br J Anaesth. 2024;133:296–304.
• 9.Bennett-Guerrero E, Romeiser JL, DeMaria S, et al; GA-CARES Trial Group. Anesthesia type during cancer surgery: results of the GA-CARES randomized, multicenter trial. Anesthesiology. 2026;144:51–62.
• 10.Enlund M, Berglund A, Enlund A, et al. Impact of general anaesthesia on breast cancer survival: a 5-year follow up of a pragmatic, randomised, controlled trial, the CAN Study, comparing propofol and sevoflurane. EClinicalMedicine. 2023;60:102037.
• 11.Cao SJ, Zhang Y, Zhang YX, et al; First Study of Perioperative Organ Protection investigators. Delirium in older patients given propofol or sevoflurane anaesthesia for major cancer surgery: a multicentre randomised trial. Br J Anaesth. 2023;131:253–265.
• 12.Sun M, Miao M, Lu Z, et al. Association of propofol vs sevoflurane maintenance anesthesia with postoperative delirium, cognitive decline, and mortality in older adults: a global comparative-effectiveness study. Crit Care. 2026;30:130.
• 13.Kampman JM, Rex S. Propofol versus sevoflurane: how many RCTs is enough? Crit Care. 2026;30:119.