PDF: STITCH
(2015)Objective
Evaluate whether early surgical evacuation improves outcomes compared to initial conservative treatment in patients with traumatic intracerebral hemorrhage (TICH).
Study Summary
Intervention
Randomized, multicenter trial (31 centers, 13 countries). Patients with TICH (≥10 mL, ≤2 clots, within 48 hrs of injury, no SDH/EDH requiring evacuation) were randomized to early surgery (within 12 hours) or initial conservative treatment. Main outcome assessed at 6 months using dichotomized Glasgow Outcome Scale (GOS).
Study Design
Arms: Array
Outcome
• 6-month mortality: 15% (surgery) vs. 33% (control); OR 0.35 (95% CI 0.16–0.75); p=0.007
• Rankin Scale and GOSE ordinal analysis: significantly better with early surgery (p=0.043 and p=0.047)
• QALY gain over 6 months: +0.019 (approx. 3.5 days)
• Subgroup with GCS 9–12 showed most benefit
Bottom Line
Early surgery for traumatic ICH significantly reduced 6-month mortality (15% vs 33%; P=0.006) vs initial conservative treatment. Primary outcome (favorable GOS) showed 10.5% absolute benefit (63% vs 53%) but did not reach significance due to premature termination at 20% of planned sample. All ordinal outcomes trended favoring surgery. No vegetative survivors in either group. First RCT for traumatic ICH.
Major Points
- Mortality halved: 15% vs 33% at 6 months (absolute difference 18.3%; P=0.006). OR 0.35 (0.16-0.75).
- Primary outcome (favorable GOS) 63% vs 53% — 10.5% absolute benefit, NS due to 20% enrollment.
- First-ever RCT of surgery for traumatic ICH. 170/840 planned patients (terminated by funder for poor UK recruitment).
- No vegetative survivors in either group — lives saved were not vegetative.
- All ordinal scales (GOS, GOSE, Rankin) showed significant trends favoring surgery (P=0.043-0.052).
- 36% of conservative patients crossed to delayed surgery — those who deteriorated had 65% unfavorable outcomes.
- GCS 9-12 subgroup: meta-analysis across STICH/STICH II/STITCH(TRAUMA) yielded OR 0.68 (0.48-0.96; P=0.03).
- 31 centers in 13 countries. India (74) and China (43) contributed most patients.
- Early surgery within median 3h of randomization, 23h of injury. 74% received surgery in ES arm.
- ~800,000 TICHs annually worldwide — confirmed benefit would impact ~84,000 patients/year.
Study Design
- Study Type
- International, multicentre, pragmatic, parallel-group, randomised controlled trial.
- Randomization
- Yes
- Blinding
- Open-label for treatment, with blinded outcome assessment.
- Sample Size
- 170
- Follow-up
- 6 months.
- Centers
- 31
- Countries
- 13 countries
Primary Outcome
Definition: Favorable outcome on the dichotomised Glasgow Outcome Scale (GOS) at 6 months (good recovery or moderate disability).
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 53% (45/85) | 63% (52/82) | 0.65 (0.35 to 1.21) | 0.171 |
Limitations & Criticisms
- The trial was stopped early, recruiting only 20% of its target sample size, which significantly limited its statistical power.
- The open-label design could introduce bias, though this was mitigated by blinded outcome assessment.
- There was a high rate of crossover from the conservative to the surgical group (36%), which could dilute the treatment effect.
- The low recruitment rate in the UK, which led to the trial's termination, raises questions about the generalizability of the findings.
Citation
Health Technol Assess 2015;19(70)