STICH II
(2013)Objective
Determine whether early surgical evacuation improves functional outcomes compared to initial conservative treatment in patients with superficial lobar intracerebral hemorrhage without intraventricular extension.
Study Summary
• In conscious patients with lobar intracerebral hemorrhage (10–100 mL, <1 cm from cortex, no IVH), early surgery did not significantly improve prognosis-based favorable outcome at 6 months compared to conservative treatment, but showed a possible survival benefit. Crossovers and heterogeneity in response may have limited the power to detect benefit.
Intervention
Randomized, multicenter trial at 78 centers in 27 countries. Patients within 48 hours of ictus, GCS motor score 5–6, eye score ≥2, randomized to early surgical evacuation (within 12 hrs) or initial conservative treatment. Follow-up at 6 months with primary outcome measured by prognosis-adjusted dichotomy of the Extended Glasgow Outcome Scale (GOSE).
Study Design
Arms: Array
Outcome
• Primary outcome (favorable GOSE): 41% (surgery) vs. 38% (control); OR 0.86 (95% CI 0.62–1.20), p=0.367
• 6-month mortality: 18% (surgery) vs. 24% (control); OR 0.71 (95% CI 0.48–1.06), p=0.095
• Modified Rankin favorable outcome: 47% vs. 44%; p=0.456
• Proportional odds analysis GOSE: OR 0.77 (95% CI 0.58–1.03), p=0.075
• Subgroup with poor prognosis showed significant benefit: OR 0.49 (95% CI 0.26–0.92), p=0.02
• 6-month mortality: 18% (surgery) vs. 24% (control); OR 0.71 (95% CI 0.48–1.06), p=0.095
• Modified Rankin favorable outcome: 47% vs. 44%; p=0.456
• Proportional odds analysis GOSE: OR 0.77 (95% CI 0.58–1.03), p=0.075
• Subgroup with poor prognosis showed significant benefit: OR 0.49 (95% CI 0.26–0.92), p=0.02
Bottom Line
Early surgery for spontaneous, superficial lobar intracerebral hemorrhage did not provide a significant overall benefit in functional outcome compared to initial conservative treatment. However, a post-hoc analysis suggested a potential survival advantage and better outcomes for patients with a poorer prognosis at presentation who underwent early surgery.
Major Points
- STICH II is the definitive trial on LOBAR ICH surgery — focused specifically on conscious patients with superficial lobar hemorrhage (≤1 cm from cortical surface) WITHOUT intraventricular hemorrhage, addressing the subgroup where surgery was thought most likely to benefit.
- 601 patients at 78 centers in 27 countries. The largest RCT of ICH surgery for lobar hemorrhage. Designed to overcome STICH I's (2005) limitation of including deep (basal ganglia/thalamic) hemorrhages where surgery is clearly futile.
- Primary outcome (prognosis-based dichotomized GOSE at 6 months) NOT significant: OR 0.86 (95% CI 0.62–1.20, p=0.367) — early surgery did NOT improve outcomes overall.
- Mortality trended lower with surgery: 18% vs 24% (p=0.095) — a 6% absolute difference that was not significant but has been widely interpreted as a signal that surgery may prevent death while converting some deaths to severe disability.
- KEY subgroup: patients with POOR predicted prognosis showed SIGNIFICANT benefit from early surgery (OR 0.49, 95% CI 0.26–0.92, p=0.02) — the one positive signal, suggesting surgery may help the sickest patients who would otherwise die.
- 21% crossover rate: 1 in 5 conservatively-treated patients required delayed surgery for neurological deterioration — this high crossover dilutes the ITT comparison and may explain why per-protocol analyses showed stronger surgical trends.
- Together with STICH I, STICH II established the clinical equipoise framework: surgery is NOT proven beneficial for most ICH, but may help lobar hemorrhages with poor prognosis and is NOT harmful — a nuanced message that supports clinical judgment over rigid protocols.
- Pragmatic design allowed surgeon choice of craniotomy technique — reflecting real-world practice variability. Modern techniques (stereotactic aspiration, endoscopic evacuation as in MISTIE III and ENRICH) may achieve hematoma reduction with less surgical morbidity.
- The prognosis-based outcome methodology was novel: each patient's expected outcome was calculated from baseline severity, and 'favorable' was defined relative to expected — attempting to detect benefit across the full severity spectrum.
- STICH II results directly informed the 2015 and 2022 AHA/ASA ICH guidelines, which state that usefulness of surgery for supratentorial ICH is 'not well established' (Class IIb) — acknowledging equipoise rather than prohibiting surgery.
Study Design
- Study Type
- International, multicentre, prospective, randomised, parallel group, pragmatic trial.
- Randomization
- Yes
- Blinding
- Open-label for treatment, with blinded outcome assessment.
- Sample Size
- 601
- Follow-up
- 6 months.
- Centers
- 78
- Countries
- 27 countries
Primary Outcome
Definition: A prognosis-based dichotomized favorable or unfavorable outcome of the 8-point Extended Glasgow Outcome Scale (GOSE) at 6 months.
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 38% favorable outcome (108/286) | 41% favorable outcome (123/297) | 0.86 (0.62 to 1.20) | 0.367 |
Limitations & Criticisms
- Overall NEGATIVE trial — primary outcome not significant (p=0.367). The positive poor-prognosis subgroup (p=0.02) was prespecified but remains hypothesis-generating given the overall null result.
- 21% crossover from conservative to delayed surgery — substantially dilutes the ITT comparison. The true surgical benefit may be larger than the ITT analysis suggests, but crossover bias works in both directions.
- Open-label treatment (PROBE design) — surgeons and treating physicians knew allocation, potentially influencing post-operative care intensity, rehabilitation referrals, and withdrawal-of-care decisions in both groups.
- Pragmatic surgical approach with no standardized technique — variation in craniotomy size, approach, hemostasis technique, and surgeon experience introduces uncontrolled heterogeneity. Modern minimally invasive approaches (endoscopic, stereotactic aspiration) were not tested.
- Excluded intraventricular hemorrhage — IVH is present in ~45% of lobar ICH cases and is an independent predictor of poor outcome. The STICH II population may not represent the full spectrum of lobar ICH.
- Excluded deep hemorrhages (basal ganglia, thalamic) — while justified by STICH I results, this means STICH II cannot inform surgical decisions for deep ICH, which represents ~60% of all ICH.
- May have been underpowered: 601 patients with an event rate lower than expected — the OR of 0.86 with wide confidence intervals (0.62–1.20) suggests that a larger trial might have detected a modest benefit.
- 48-hour enrollment window — some patients may have had hematoma stabilization by the time of surgery, reducing the potential benefit of evacuation. Modern protocols emphasize ultra-early surgery within 6–8 hours.
- Prognosis-based dichotomized outcome is a non-standard endpoint — unfamiliar to many readers and makes comparison to other ICH trials using standard mRS challenging.
Citation
Lancet 2013; 382: 397-408