Skip to main content

Foundational Trials

STOP-IT

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Trial of Short-Course Antimicrobial Therapy for Intraabdominal Infection
  • Acronym: STOP-IT (Study to Optimize Peritoneal Infection Therapy)
  • Year: 2015
  • Journal published in: The New England Journal of Medicine
  • Citation: Sawyer RG, Claridge JA, Nathens AB, et al. Trial of Short-Course Antimicrobial Therapy for Intraabdominal Infection. N Engl J Med. 2015;372(21):1996-2005.

Context & Rationale

  • Background
    Complicated intra-abdominal infection (cIAI) is common, morbid, and historically treated with prolonged antimicrobial courses that were frequently continued until “physiological recovery” (resolution of fever/leukocytosis and return of gut function), despite limited trial evidence.
    Guidelines before STOP-IT emphasised that once adequate source control is achieved, shorter antimicrobial courses (typically 4–7 days) are usually sufficient, yet practice variation and prolonged exposure persisted.2
  • Research Question/Hypothesis
    After adequate source control for cIAI, a fixed short course of antimicrobials (4 days) would yield similar 30-day clinical outcomes to a longer course guided by resolution of physiological abnormalities.
  • Why This Matters
    If short-course therapy is safe after source control, it reduces antimicrobial exposure (and its collateral damage) without increasing recurrent infection, surgical-site infection, or death, reinforcing source control as the dominant determinant of outcome and enabling stewardship in surgical/critical care pathways.

Design & Methods

  • Research Question: In adults with cIAI and adequate source control, does fixed-duration antimicrobial therapy (4 days) achieve similar 30-day outcomes compared with antimicrobials continued until 2 days after resolution of fever, leukocytosis, and gastrointestinal dysfunction (maximum 10 days)?
  • Study Type: Randomised, multicentre, investigator-initiated, pragmatic, open-label clinical trial; allocation stratified by enrolling site and appendiceal vs non-appendiceal source; source control achieved by open, laparoscopic, or percutaneous intervention.
  • Population:
    • Setting: Hospital-based surgical/percutaneous source-control pathways (multiple centres).
    • Inclusion criteria (protocol-defined): Age ≥16 years; ability to consent; intra-abdominal infection requiring hospitalisation and managed with open/laparoscopic/percutaneous intervention; within 24 hours of initial intervention had WBC >11,000/mm3, temperature ≥38.0°C, or gastrointestinal dysfunction preventing normal diet intake; adequate source control per local investigator and principal investigator.
    • Key exclusions (protocol-defined): Inability to obtain consent; inadequate source control; high likelihood of death within 72 hours; planned relaparotomy; perforated gastric/duodenal ulcer treated within 24 hours of symptom onset; traumatic (including iatrogenic/intra-operative) bowel injury treated within 12 hours; non-perforated/non-gangrenous appendicitis or cholecystitis; gangrenous appendicitis or peritonitis without confirmatory cultures or with negative cultures; ischaemic/necrotic intestine without perforation and without positive cultures; intra-abdominal infection with active necrotising pancreatitis (discrete pancreatic abscess/infected pseudocyst permissible); primary spontaneous bacterial peritonitis; continuous ambulatory peritoneal dialysis catheter-related infection; primary skin closure of an open incision in diffuse peritonitis (open abdomen scenarios); pregnancy; concurrent infection requiring >4 days of therapy; enrolment in another therapeutic trial.
  • Intervention:
    • Fixed short-course arm: Antimicrobials stopped within 24 hours after assignment to the 4-day strategy (protocol operationalisation: total of 3–4 full days considered protocol-adherent); antimicrobial choice otherwise per treating team.
  • Comparison:
    • Physiological-resolution arm: Antimicrobials continued until 2 ± 1 days after resolution of leukocytosis (WBC ≤11,000/mm3), fever (temperature <38°C), and gastrointestinal dysfunction adequate to allow enteral nutrition (maximum total duration 10 days).
  • Blinding: Unblinded (open-label); outcomes were largely objective (infection recurrence, surgical-site infection, death), but clinical thresholds for re-imaging and sampling could plausibly be influenced by knowledge of treatment duration.
  • Statistics: Protocol power calculation: 1,120 patients planned to detect a 10% absolute difference in the primary composite outcome with 80% power at a two-sided 5% significance level; primary analyses reported as intention-to-treat with effect estimates presented using between-group differences and 95% confidence intervals.
  • Follow-Up Period: 30 days after the index source-control procedure (post-discharge outcomes obtained via patient/physician contact and/or medical record review).

Key Results

This trial was not reported as stopped early. The protocol planned 1,120 patients, whereas 518 were randomised in the published report.

Outcome Fixed duration (≈4 days) Physiological resolution (≤10 days) Effect p value / 95% CI Notes
Primary composite (SSI, recurrent intra-abdominal infection, or death by 30 days) 56/257 (21.8%) 58/260 (22.3%) Δ −0.5 percentage points 95% CI −7.0 to 8.0; P=0.92 No clinically important signal favouring longer therapy within observed precision.
Surgical-site infection 19/257 (7.4%) 17/260 (6.5%) Δ 0.9 percentage points 95% CI −3.8 to 5.6; P=0.74 Time to diagnosis (among those diagnosed): 8.8 ± 0.4 vs 15.1 ± 0.6 days; P<0.001.
Recurrent intra-abdominal infection 40/257 (15.6%) 45/260 (17.3%) Δ −1.7 percentage points 95% CI −8.4 to 5.0; P=0.60 Time to diagnosis (among those diagnosed): 10.8 ± 0.4 vs 15.1 ± 0.5 days; P<0.001.
Death by 30 days 1/257 (0.4%) 2/260 (0.8%) Δ −0.4 percentage points 95% CI −3.1 to 2.2; P=0.56 Very low mortality; limits power to detect small mortality differences.
Total duration of antimicrobial therapy Median 4.0 days (IQR 4.0–5.0) Median 8.0 days (IQR 5.0–10.0) Δ −4.0 days 95% CI −4.7 to −3.3; P<0.001 Clear separation of the exposure of interest.
Antimicrobial-free days at 30 days Median 25 (IQR 21–26) Median 21 (IQR 18–25) Δ 4 days 95% CI 2.9 to 4.7; P<0.001 Stewardship-relevant benefit without worsening the primary outcome.
Extra-abdominal infection (any) 23/257 (8.9%) 13/260 (5.0%) Δ 3.9 percentage points 95% CI −0.8 to 8.5; P=0.11 Numerically higher in fixed-duration arm; not statistically significant.
Clostridioides difficile infection 5/257 (1.9%) 3/260 (1.2%) Δ 0.7 percentage points 95% CI −1.5 to 2.9; P=0.71 Low event rate; wide uncertainty.
Hospital-free days at 30 days Median 22 (IQR 16–26) Median 23 (IQR 18–26) Δ −1 day 95% CI −2.0 to 0.0; P=0.22 No signal that longer therapy shortened hospitalisation.
  • The primary composite outcome was similar: 21.8% with ≈4 days vs 22.3% with physiology-guided duration (Δ −0.5 percentage points; 95% CI −7.0 to 8.0; P=0.92).
  • Exposure separation was large: median antimicrobial duration 4.0 (IQR 4–5) vs 8.0 days (IQR 5–10) (Δ −4.0 days; 95% CI −4.7 to −3.3; P<0.001).
  • Diagnosis of intra-abdominal complications occurred later in the longer-duration group (e.g., recurrent intra-abdominal infection: 10.8 ± 0.4 vs 15.1 ± 0.5 days; P<0.001), consistent with “suppression” rather than prevention of failure.

Internal Validity

  • Randomisation and allocation: Centralised randomisation with stratification by site and appendiceal vs non-appendiceal source supports allocation concealment and balance.
  • Attrition and exclusions: 518 randomised; 517 included in the 30-day analysis (≈99.8% follow-up), with minimal post-randomisation loss, limiting attrition bias.
  • Performance/detection bias: Open-label design introduces potential for differential re-imaging/sampling thresholds and clinician behaviour, particularly for softer secondary outcomes (e.g., extra-abdominal infections), though the primary composite components are relatively objective.
  • Protocol adherence: Protocol-adherent antimicrobial duration was achieved in 81.8% of the fixed-duration group and 72.7% of the physiological-resolution group, indicating meaningful but incomplete fidelity; nonadherence plausibly biases toward null differences.
  • Baseline comparability: Groups were well balanced (e.g., APACHE II 9.9 ± 3.9 vs 10.3 ± 4.0; healthcare-associated infection 36.8% vs 36.2%; percutaneous drainage 32.6% vs 33.1%).
  • Heterogeneity: Broad anatomical sources (colon/rectum 34.1% vs 31.5%; small intestine 17.8% vs 15.0%; biliary 7.8% vs 9.2%) increases clinical relevance but may dilute effects if duration matters only in specific phenotypes; prespecified subgroup analyses did not identify a clear differential effect.
  • Timing: Protocol operationalisation delayed group assignment until day 4 after the index intervention (with prior “common” early therapy), meaning the randomised contrast chiefly tested “stop at ~day 4” vs “continue beyond day 4”; this is appropriate for the clinical decision point but selects for early survivors with adequate source control.
  • Dose and separation of the variable of interest: Clear separation was achieved: median therapy 4.0 (IQR 4–5) vs 8.0 days (IQR 5–10); antimicrobial-free days at 30 were 25 (IQR 21–26) vs 21 (IQR 18–25).
  • Outcome assessment: Primary outcomes were clinically defined and time-limited (30 days), but ascertainment of recurrent infection and extra-abdominal infections depends on clinician-initiated diagnostics.
  • Statistical rigour: The published sample (518) was substantially below the protocol target (1,120), constraining precision for a formal equivalence/non-inferiority claim; nevertheless, point estimates for the primary composite were near-identical.

Conclusion on Internal Validity: Overall, internal validity appears moderate: randomisation and near-complete follow-up were strong, exposure separation was large, but open-label delivery and incomplete protocol adherence, plus a smaller-than-planned sample for an equivalence question, limit inferential strength for small differences.

External Validity

  • Population representativeness: Adults with a wide range of cIAI sources treated via laparotomy, laparoscopy, or percutaneous drainage; moderate illness severity (APACHE II ≈10) and low mortality.
  • Key exclusions shaping generalisability: Patients without “adequate source control”, those anticipated to die within 72 hours, planned relaparotomy/open abdomen scenarios, primary spontaneous bacterial peritonitis, and active necrotising pancreatitis were excluded.
  • Applicability across systems: Findings translate best to systems able to deliver timely, effective source control and close clinical reassessment; settings with delayed/imperfect source control, high resistance burden, or limited follow-up infrastructure may see different risk–benefit balances.
  • Subpopulations: Evidence is strongest for typical postoperative/percutaneous-drainage cIAI with adequate source control; less informative for profoundly immunocompromised hosts, ongoing leaks, or open abdomen strategies (excluded by protocol design).

Conclusion on External Validity: Overall generalisability is good for adults with cIAI who receive adequate source control and are managed in similar multidisciplinary hospital pathways, but limited for patients with ongoing contamination, planned re-operations, or pancreatitis-related necrosis.

Strengths & Limitations

  • Strengths: Pragmatic multicentre randomised design; clinically meaningful stewardship question at the real decision point (stop vs continue beyond day 4); excellent follow-up completeness; substantial treatment separation (median 4 vs 8 days) with robust, patient-centred outcomes at 30 days.
  • Limitations: Open-label delivery; protocol-adherence was imperfect (81.8% vs 72.7%); enrolment below protocol sample-size target for an equivalence framing; “adequate source control” was investigator-judged and excludes key high-risk phenotypes (open abdomen/planned relaparotomy), potentially underestimating risks in those groups.

Interpretation & Why It Matters

  • Practical clinical implication
    After adequate source control for cIAI, stopping antimicrobials at ~4 days produced similar 30-day infection/death outcomes to continuing until clinical resolution (21.8% vs 22.3%), while reducing antibiotic exposure by a median of 4 days (4 vs 8) and increasing antimicrobial-free days (25 vs 21).
  • Mechanistic insight
    Longer therapy did not prevent failures and was associated with later recognition of recurrent infection/SSI (e.g., recurrent infection diagnosis 15.1 ± 0.5 vs 10.8 ± 0.4 days), consistent with suppression rather than elimination of an inadequately controlled source.
  • How it shifts practice
    STOP-IT operationalised a stewardship principle: once source control is achieved, physiological abnormalities should prompt reassessment for uncontrolled infection rather than automatic extension of broad-spectrum antimicrobials.

Controversies & Subsequent Evidence

  • Equivalence vs “no difference”: The protocol was framed to detect a 10% absolute difference with 1,120 patients, but the published sample was 518; therefore, the trial is best interpreted as showing closely similar point estimates with confidence intervals that do not exclude modest differences, rather than establishing formal equivalence/non-inferiority.
  • Source control as the real intervention: The accompanying editorial emphasised that cIAI outcomes are dominated by adequacy of source control and that prolonged antibiotics can distract from timely recognition of ongoing sepsis drivers (e.g., undrained collections or leaks).1
  • Outcome ascertainment in an open-label trial: Clinician thresholds for re-imaging and culture acquisition can influence diagnosis timing; the observed later diagnosis of SSI/recurrent infection in the longer-duration group may reflect delayed detection due to ongoing antibiotic exposure.
  • Background guideline alignment: Pre-STOP-IT guidance already suggested that short courses (often 4–7 days) are sufficient after adequate source control, and STOP-IT provided randomised evidence supporting a fixed short-course strategy in appropriately selected patients.2
  • Observational context: Large international observational data demonstrate substantial morbidity in cIAI and highlight practice variation that makes stewardship-relevant duration trials clinically consequential.3

Summary

  • In adults with cIAI and investigator-judged adequate source control, a fixed ~4-day antimicrobial course produced similar 30-day outcomes to a physiology-guided longer course (21.8% vs 22.3%; Δ −0.5 percentage points; 95% CI −7.0 to 8.0; P=0.92).
  • Antibiotic exposure was materially reduced (median 4.0 vs 8.0 days; Δ −4.0 days; 95% CI −4.7 to −3.3; P<0.001) with more antimicrobial-free days (25 vs 21; P<0.001).
  • Longer therapy did not reduce recurrent intra-abdominal infection (15.6% vs 17.3%; P=0.60) or surgical-site infection (7.4% vs 6.5%; P=0.74).
  • Diagnosis of recurrent infection and SSI occurred later in the longer-duration group (P<0.001 for both), supporting reassessment for source-control failure rather than default antibiotic extension.
  • Generalisability is strongest for patients with effective source control; excluded phenotypes (open abdomen/planned relaparotomy, pancreatitis necrosis, anticipated early death) remain uncertain.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • “Not reported” items in Further Reading reflect limitations of the provided source set for identifying and fully verifying additional eligible published studies with DOI-linked citations.

Overall Takeaway

STOP-IT challenged the entrenched practice of continuing antimicrobials until physiological recovery after cIAI source control. In a pragmatic multicentre randomised trial with clear exposure separation (median 4 vs 8 days), fixed short-course therapy yielded similar 30-day outcomes and materially reduced antibiotic exposure, cementing a source-control-first, stewardship-aligned approach to postoperative abdominal sepsis.

Overall Summary

  • After adequate source control for cIAI, ~4 days of antimicrobials achieved similar 30-day outcomes to longer, physiology-guided therapy while reducing antibiotic exposure by a median of 4 days.

Bibliography