Publication

• Title: Selective decontamination of the digestive tract during ventilation in the ICU
• Acronym: SuDDICU
• Year: 2025
• Journal published in: The New England Journal of Medicine
• Citation: SuDDICU Investigators for the Australia and New Zealand Intensive Care Society Clinical Trials Group and the Canadian Critical Care Trials Group. Selective decontamination of the digestive tract during ventilation in the ICU. N Engl J Med. 2025;epublished October 29th

Context & Rationale

• Background

  • Selective decontamination of the digestive tract (SDD) combines topical nonabsorbable antimicrobials applied to the oropharynx and gastrointestinal tract (often with a short systemic antibiotic course) to suppress potentially pathogenic flora and prevent ICU-acquired infection.
  • Earlier trial programmes (predominantly in low-resistance ICUs) suggested reductions in ventilator-associated pneumonia and bloodstream infection, with meta-analytic signals for reduced mortality.
  • International uptake has been constrained by (i) uncertainty about reproducibility/generalisation outside those settings and (ii) ecological concerns that prophylactic broad antibiotic exposure could increase antimicrobial resistance at unit or health-system level.
  • A prior large pragmatic cluster cross-over trial (SuDDICU Australia) did not show a statistically significant reduction in hospital mortality, maintaining equipoise for a mortality effect in contemporary practice.¹
  • A Bayesian meta-analysis integrating contemporary trials estimated a high posterior probability of mortality benefit, emphasising the sensitivity of inference to heterogeneity and modelling assumptions.²
  • Evidence focusing on antimicrobial resistance has remained mixed and incomplete, especially for long-term, unit-level outcomes relevant to stewardship decisions.³
• Research Question/Hypothesis

  • Patient-level: In invasively ventilated ICU patients, does SDD (vs standard care) reduce in-hospital mortality within 90 days?
  • Unit-level safety: Is implementing SDD non-inferior (i.e., does not meaningfully worsen) ICU-level incidence of antibiotic-resistant organisms, bloodstream infection, or Clostridioides difficile infection?
• Why This Matters

  • If SDD confers even a modest absolute mortality reduction, its potential population benefit is substantial given the global burden of mechanical ventilation.
  • Conversely, widespread prophylactic antibiotic exposure could impose externalities (selection pressure, resistance propagation) that may not be captured by individual-level endpoints.
  • Embedding an ecological (unit-level) surveillance framework alongside patient outcomes directly addresses the central implementation dilemma: net benefit vs net harm at ICU population level.

Design & Methods

• Research Question: Among adults receivingisonrally ventilated in participating ICUs, does implementing SDD (vs standard care) reduce in-hospital mortality within 90 days while remaining ecologically safe with respect to antibiotic resistance and key healthcare-associated infections?
• Study Type: Pragmatic, multicentre, international (Australia and Canada), cluster-randomised, cross-over trial with two 12-month intervention periods separated by a 3-month inter-period gap; bedside care unblinded.
• Population:
  • Setting: 26 adult ICUs (Australia and Canada) implementing SDD or standard care according to randomised sequence; UK sites suspended recruitment and were not included in the primary report.
  • Inclusion (patient-level): Adults (≥16 years) receiving invasive mechanical ventilation, expected to remain ventilated ≥48 hours.
  • Key exclusions: Expected death within 12 hours; previous enrolment; known allergy/intolerance to study drugs; pregnancy; other protocol-specified contraindications to study interventions.
  • Enrolment: 9,289 patients enrolled (SDD period n=4,223; standard-care period n=5,066).
• Intervention:
  • SDD regimen: topical oropharyngeal paste plus gastric suspension containing colistin, tobramycin, and nystatin administered at regular intervals while intubated.
  • Systemic component: a short (4-day) systemic antibiotic course at SDD initiation (third-generation cephalosporin regimen; country-specific with protocolised alternatives for allergy).
  • Stopping rules: intervention stopped at extubation, ICU discharge, or if protocol-defined cessation criteria were met.
• Comparison:
  • Standard care: no topical SDD paste/suspension and no mandated prophylactic systemic antibiotic; all other care (including local infection-prevention bundles and clinician-directed antimicrobials for suspected/confirmed infection) per usual practice.
• Blinding: Unblinded (cluster-level implementation makes blinding impractical); primary outcome (death) objective, but infection/microbiology outcomes potentially influenced by diagnostic sampling and antibiotic prescribing behaviour.
• Statistics: Originally, 6,000 patients were required to detect a 4.2% absolute reduction in hospital mortality (from 29.0% to 24.8%) with 80% power at a two-sided 5% significance level; the target sample size was later increased (to 10,000 and then 15,000) because of higher observed intracluster correlation and recruitment disruptions. Primary analyses were intention-to-treat at patient-level using models accounting for clustering and period; secondary clinical outcomes were multiplicity-adjusted (Holm step-down) in the prespecified hierarchy.
• Follow-Up Period: In-hospital mortality assessed up to 90 days after enrolment; clinical secondary outcomes assessed through hospitalisation (and ICU course) as specified; antibiotic utilisation assessed over days 1–28; ecological outcomes assessed at ICU level across pre-trial, intervention, inter-period, and post-trial surveillance epochs.

Key Results

This trial was not stopped early. One formal interim analysis was conducted after completion of the first Australian intervention period, and the trial continued without protocol changes; recruitment was later disrupted by the COVID-19 pandemic and UK sites suspended recruitment (UK data not included in the primary report).

• Mortality: No statistically significant difference in the primary endpoint; the point estimate favoured SDD, but the 95% CI included both modest benefit and no benefit (OR 0.95; 95% CI 0.86 to 1.05).
• Mechanistic/clinical signals: Patient-level bloodstream infection (3.6% vs 5.7%) and new antibiotic-resistant organism culture (16.8% vs 26.8%) were lower during SDD periods, while overall antibiotic exposure was higher (2.0 ± 2.0 vs 1.6 ± 1.9 defined daily doses).
• Ecological uncertainty: Unit-level non-inferiority for antibiotic-resistant organism culture incidence was not demonstrated (upper CI exceeded the +2% margin), so the trial could not exclude a clinically important ecological harm despite point estimates favouring SDD.

Internal Validity

• Randomisation and allocation: 26 ICUs were randomised to intervention sequence (cluster cross-over), reducing between-unit confounding relative to parallel cluster designs; allocation concealment at patient level is not applicable because assignment was determined by period-level ICU implementation.
• Screening/enrolment and selection bias risk: Of 12,784 screened during intervention periods, 9,289 were enrolled; “not enrolled for other/unknown reasons” accounted for 2,294/4,144 (55.4%) and 695/2,447 (28.4%) in SDD periods vs 2,423/4,175 (58.0%) and 444/2,018 (22.0%) in standard-care periods, creating potential for differential enrolment by period.
• Dropout/exclusions: Primary outcome data were available for 4,215/4,223 (99.8%) in SDD periods and 5,065/5,066 (>99.9%) in standard-care periods; consent withdrawal affecting intervention delivery was uncommon (supplementary adherence tables report 6 [0.2%] withdrawals impacting oral paste delivery and 5 [0.1%] for gastric suspension).
• Performance/detection bias: Open-label delivery could influence culture acquisition, antibiotic prescribing, and diagnostic labelling; primary outcome is objective, but microbiology endpoints depend on clinical sampling practices (partly mitigated for ecological outcomes by protocolised surveillance sampling windows).
• Protocol adherence: Supplementary protocol deviation reporting showed delayed initiation (neither oral paste nor gastric suspension started within 6 hours) in 710/4,233 (16.8%) and at least one full day without SDD-compliant systemic antibiotic in 620/4,233 (14.7%); clinician refusal was reported for 121/4,233 (2.9%).
• Baseline characteristics: Broad comparability, but some imbalances existed (e.g., median APACHE III score 68.0 [IQR 49.0–89.0] in SDD vs 73.0 [53.0–95.0] in standard care; median time from ICU admission to enrolment 14.3 vs 6.2 hours); adjusted analysis did not materially change conclusions.
• Heterogeneity: Cluster cross-over mitigates stable between-ICU differences, but time-varying practice changes (including pandemic-era pressures) remain a plausible contributor to residual heterogeneity; country-level and period-level effects are inherently intertwined in this design.
• Timing and dose: The intervention was delivered during mechanical ventilation; initiation was not uniformly within 6 hours; systemic antibiotic exposure was substantially higher in SDD periods (84.2% vs 58.8% received an SDD-compliant systemic antibiotic), demonstrating separation but also introducing stewardship-relevant co-intervention differences.
• Separation of the variable of interest (numeric):
  • Defined daily doses of antibiotics (days 1–28): 2.0 ± 2.0 (SDD) vs 1.6 ± 1.9 (standard care); mean difference 0.3 (95% CI 0.2 to 0.4).
  • New antibiotic-resistant organism cultured (patient-level): 16.8% vs 26.8%; difference −9.6 percentage points (95% CI −12.3 to −6.9).
  • New bloodstream infection: 3.6% vs 5.7%; difference −2.1 percentage points (95% CI −3.2 to −0.9).
• Outcome assessment and statistical rigour: Prespecified modelling accounted for cluster and period; multiplicity correction was used for secondary clinical outcomes; achieved sample size (9,289) was below the ultimately planned target, with implications for power to detect small mortality differences.

Conclusion on Internal Validity: Internal validity appears moderate: randomised cluster cross-over design and near-complete primary outcome ascertainment support credible mortality inference, but open-label implementation, substantial “not enrolled/unknown” screening attrition, and imperfect protocol adherence introduce risks of selection and detection bias for non-mortality endpoints.

External Validity

• Population representativeness: Broad adult ventilated ICU population in high-income systems (Australia/Canada), with pragmatic inclusion and relatively few absolute exclusions; results most applicable to mixed medical-surgical ICUs with similar baseline resistance ecology.
• Healthcare system and stewardship context: Implementation requires reliable access to compounded topical agents and the capacity for ongoing microbiology surveillance; the trial setting likely reflects comparatively lower baseline resistance than many global ICUs, which may alter risk–benefit balance.
• Applicability to higher-resistance settings: Unit-level non-inferiority for antibiotic-resistant organism incidence was not demonstrated, so extrapolation to ICUs with higher baseline resistance or different dominant pathogens is particularly uncertain.
• Implementation fidelity in routine practice: The observed delays and missed doses (e.g., 16.8% not started within 6 hours; 14.7% with at least one full day without SDD-compliant systemic antibiotic) suggest that “real-world” adoption may deliver variable exposure, potentially attenuating benefit and complicating stewardship impact assessment.

Conclusion on External Validity: External validity is moderate: findings generalise well to comparable Australasian/Canadian mixed ICUs, but translation to settings with higher antimicrobial resistance prevalence, different stewardship constraints, or limited surveillance capacity is uncertain.

Strengths & Limitations

• Strengths:
  • Large, pragmatic, multinational cluster cross-over design with 9,289 enrolled patients.
  • Hard primary endpoint (in-hospital death within 90 days) with minimal missingness.
  • Integrated ecological surveillance framework explicitly addressing resistance and infection externalities.
  • Prespecified statistical modelling accounting for clustering/period effects and multiplicity-adjusted secondary outcomes.
• Limitations:
  • Open-label bedside implementation introduces potential performance and detection bias for culture-dependent outcomes.
  • Substantial screening attrition labelled as “not enrolled for other/unknown reasons,” with risk of selection bias and period-dependent enrolment differences.
  • Protocol adherence was imperfect (e.g., delayed initiation and missed doses), potentially diluting effect estimates and complicating interpretation of “true” biological efficacy.
  • Achieved sample size was below the ultimately planned target, limiting power to detect small absolute mortality differences.
  • Ecological non-inferiority for antibiotic-resistant organism incidence was not demonstrated, leaving residual uncertainty about population-level safety despite favourable point estimates.

Interpretation & Why It Matters

• Clinical practice signal

  Routine SDD implementation in Australia/Canada did not significantly reduce in-hospital mortality (OR 0.95; 95% CI 0.86 to 1.05), despite reductions in bloodstream infection and patient-level resistant organism cultures.
• Stewardship trade-off

  Antibiotic exposure increased (defined daily doses 2.0 ± 2.0 vs 1.6 ± 1.9; mean difference 0.3), while ecological non-inferiority for unit-level antibiotic-resistant organism culture incidence was not confirmed—exactly the tension stewardship programmes must adjudicate.
• Research implications

  The mortality effect estimate narrows the plausible benefit in contemporary practice and strengthens the case that future work should focus on (i) identifying patient subgroups most likely to benefit and (ii) long-horizon, system-level resistance consequences under sustained implementation.

Controversies & Subsequent Evidence

• Mortality uncertainty vs meta-analytic belief: The trial’s confidence interval includes clinically meaningful benefit and no benefit; this remains in tension with Bayesian meta-analytic conclusions that assign a high probability of mortality benefit when integrating contemporary trials.²
• Resistance externalities are not “settled” by patient-level cultures: Patient-level resistant organism cultures were lower during SDD periods, but unit-level non-inferiority for resistant organism incidence was not demonstrated; this aligns with prior systematic review findings that resistance outcomes are heterogeneous and context-dependent.³
• Selection and ascertainment bias remain live critiques: Differential non-enrolment for “other/unknown reasons” and open-label influence on culture acquisition could bias culture-dependent endpoints (favourably or unfavourably), complicating causal interpretation of microbiology outcomes independent of mortality.
• How SuDDICU (NEJM 2025) sits with SuDDICU Australia (JAMA 2022): The two large pragmatic trials now show broadly consistent mortality point estimates with overlapping uncertainty, making very large mortality effects less plausible in contemporary high-income practice.¹
• Safety debate framing: The methodological difficulty of “proving” ecological safety (given low incidence outcomes, long causal horizons, and spillovers) remains a central objection in the SDD literature, and is not fully resolved by short-horizon non-inferiority testing.⁴

Summary

• Large multinational (Australia/Canada) cluster cross-over trial of SDD in mechanically ventilated ICU patients (n=9,289).
• No statistically significant reduction in 90-day in-hospital mortality (27.0% vs 28.2%; OR 0.95; 95% CI 0.86 to 1.05).
• Lower patient-level bloodstream infection (3.6% vs 5.7%) and new antibiotic-resistant organism cultures (16.8% vs 26.8%), alongside higher antibiotic exposure (defined daily doses 2.0 ± 2.0 vs 1.6 ± 1.9).
• Ecological non-inferiority for unit-level antibiotic-resistant organism incidence was not demonstrated, so population-level safety remains uncertain.
• Internal validity for mortality is supported by objective outcome and near-complete follow-up, but open-label implementation, non-enrolment, and recall/diagnostic biases limit inference for microbiology endpoints.

Further Reading

Other Trials

• 2022Hammond NE, Myburgh J, Seppelt I, et al. Effect of selective decontamination of the digestive tract on hospital mortality in critically ill patients receiving mechanical ventilation: the SuDDICU randomized clinical trial. JAMA. 2022;328(19):1911-1921.
• 2009de Smet AMGA, Kluytmans JAJW, Cooper BS, et al. Decontamination of the digestive tract and oropharynx in ICU patients. N Engl J Med. 2009;360(1):20-31.
• 2003de Jonge E, Schultz MJ, Spanjaard L, et al. Effects of selective decontamination of digestive tract on mortality and acquisition of resistant bacteria in intensive care: a randomised controlled trial. Lancet. 2003;362:1011-1016.
• 1992Gastinne H, Wolff M, Delatour F, et al. A controlled trial in intensive care units of selective decontamination of the digestive tract with nonabsorbable antibiotics. N Engl J Med. 1992;326:594-599.
• 2014Oostdijk EAN, Kesecioglu J, Schultz MJ, et al. Effects of decontamination of the oropharynx and intestinal tract on antibiotic resistance in ICUs: a randomized clinical trial. JAMA. 2014;312:1429-1437.

Systematic Review & Meta Analysis

• 2022Hammond NE, Finfer S, Delaney A, et al. Association between selective decontamination of the digestive tract and in-hospital mortality in critically ill patients receiving mechanical ventilation: a systematic review and meta-analysis. JAMA. 2022;328(19):1922-1934.
• 2013Daneman N, Sarwar S, Fowler RA, Cuthbertson BH. Effect of selective decontamination on antimicrobial resistance in intensive care units: a systematic review and meta-analysis. Lancet Infect Dis. 2013;13(4):328-341.
• 2014Price R, MacLennan G, Glen J. Selective digestive or oropharyngeal decontamination and topical oropharyngeal chlorhexidine for prevention of death in general intensive care: systematic review and network meta-analysis. BMJ. 2014;348:g2197.
• 2021Minozzi S, Pifferi S, Brazzi L, et al. Topical antibiotic prophylaxis to reduce respiratory tract infections and mortality in adults receiving mechanical ventilation. Cochrane Database Syst Rev. 2021;1:CD000022.

Observational Studies

• 2014Canter RR, Harvey SE, Harrison DA, Campbell MK, Rowan KM, Cuthbertson BH. Observational study of current use of selective decontamination of the digestive tract in UK critical care units. Br J Anaesth. 2014;113:610-617.
• 2018Sánchez-Ramírez C, Hípola-Escalada S, Cabrera-Santana M, et al. Long-term effect of selective decontamination of the digestive tract on antimicrobial resistance in an intensive care unit with high endemicity. Crit Care. 2018.
• 2014Houben AJ, Oostdijk EAN, van der Voort PHJ, et al. Selective decontamination of the digestive tract and the incidence of antimicrobial resistance: ecological observations from intensive care. J Antimicrob Chemother. 2014.
• 2024Rodríguez-Gascón A, Fernández-Sainz I, Isla A, et al. Microbiological changes following selective decontamination of the digestive tract: a long-term comparison in intensive care. Eur J Clin Microbiol Infect Dis. 2024.
• 2025Fernández-Fernández D, Escario-Pinel M, González de Molina FJ, et al. Effect of implementation of selective digestive decontamination in a medical-surgical ICU: a before-and-after study. J Crit Care Med (Targu Mures). 2025.

Guidelines

• 2025Rosenthal VD, Bat-Erdene I, Gupta D, et al. Strategies to prevent ventilator-associated pneumonia: international position statement. Int J Infect Dis. 2025;151:413-429.
• 2022Klompas M, Branson R, Eichenwald EC, et al. Strategies to prevent ventilator-associated pneumonia, ventilator-associated events, and nonventilator hospital-acquired pneumonia in acute-care hospitals: 2022 update. Infect Control Hosp Epidemiol. 2022.
• 2017Torres A, Niederman MS, Chastre J, et al. International ERS/ESICM/ESCMID/ALAT guidelines for the management of hospital-acquired pneumonia and ventilator-associated pneumonia. Eur Respir J. 2017;50:1700582.
• 2016Kalil AC, Metersky ML, Klompas M, et al. Management of adults with hospital-acquired and ventilator-associated pneumonia: 2016 clinical practice guidelines by the IDSA and ATS. Clin Infect Dis. 2016;63(5):e61-e111.
• 2005American Thoracic Society; Infectious Diseases Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med. 2005;171(4):388-416.

Notes

• Effect measures in the results table reflect those reported in the manuscript/supplement (odds ratios, absolute percentage-point differences, and mean differences); no additional calculations were performed.
• Culture-based endpoints (including “new antibiotic-resistant organism cultured”) are intrinsically sensitive to sampling intensity and diagnostic thresholds in open-label pragmatic ICU trials.

Overall Takeaway

In a large pragmatic cluster cross-over trial spanning Australia and Canada, implementing SDD during mechanical ventilation did not significantly reduce 90-day in-hospital mortality, despite favourable signals for bloodstream infection and patient-level resistant organism cultures. However, the ecological non-inferiority analysis for unit-level resistant organism incidence was inconclusive, leaving the stewardship trade-off unresolved and positioning SDD as a strategy whose adoption depends heavily on local resistance ecology, surveillance capability, and tolerance for residual uncertainty.

Bibliography

• 1.Hammond NE, Myburgh J, Seppelt I, et al. Effect of selective decontamination of the digestive tract on hospital mortality in critically ill patients receiving mechanical ventilation: the SuDDICU randomized clinical trial. JAMA. 2022;328(19):1911-1921.
• 2.Hammond NE, Finfer S, Delaney A, et al. Association between selective decontamination of the digestive tract and in-hospital mortality in critically ill patients receiving mechanical ventilation: a systematic review and meta-analysis. JAMA. 2022;328(19):1922-1934.
• 3.Daneman N, Sarwar S, Fowler RA, Cuthbertson BH. Effect of selective decontamination on antimicrobial resistance in intensive care units: a systematic review and meta-analysis. Lancet Infect Dis. 2013;13(4):328-341.
• 4.Hurley JC. Establishing the safety of selective decontamination of the digestive tract in intensive care: a conversation at the margin of current knowledge. Trials. 2023;24:337.