PDF: IHAST
(2005)Objective
To determine whether intraoperative cooling during open craniotomy would improve the outcome among patients with acute aneurysmal subarachnoid hemorrhage.
Study Summary
• Good outcome (Glasgow Outcome Score of 1) in 66% of hypothermia group vs 63% of normothermia group (OR 1.14, 95% CI 0.88-1.48, p=0.32)
• Postoperative bacteremia was more common in hypothermia group (5% vs 3%, p=0.05)
Intervention
Patients randomly assigned to intraoperative hypothermia (target 33°C using surface cooling) or normothermia (target 36.5°C) during surgical aneurysm clipping
Inclusion Criteria
Age ≥18 years, acute nontraumatic SAH from radiologically demonstrated intracranial aneurysm within 14 days, World Federation of Neurological Surgeons score of I, II, or III (good-grade), Rankin score 0-1 before hemorrhage
Study Design
Arms: Hypothermia (33°C) vs Normothermia (36.5°C)
Patients per Arm: 499 hypothermia, 501 normothermia
Outcome
• Death at follow-up: 6% in both groups
• No significant differences in ICU stay, hospital length, or discharge destination
Bottom Line
Intraoperative hypothermia (33°C) did not improve neurologic outcomes after craniotomy among good-grade patients (WFNS I, II, or III) with aneurysmal subarachnoid hemorrhage. At 90-day follow-up, there were no significant differences between hypothermia and normothermia groups in rates of good outcome (66% vs 63%, OR 1.14, 95% CI 0.88-1.48, p=0.32), death (6% in both groups), ICU length of stay, or total hospitalization duration. Postoperative bacteremia was more common in the hypothermia group (5% vs 3%, p=0.05). The trial's 95% confidence intervals rule out the clinically meaningful 10% absolute improvement in good outcomes that the study was designed to detect.
Major Points
- Multicenter prospective randomized partially blinded trial at 30 centers comparing intraoperative hypothermia vs normothermia
- 1001 patients randomized from February 2000 to April 2003; 1000 patients with follow-up data
- Patients had good-grade SAH (WFNS I, II, or III) and underwent surgery within 14 days of hemorrhage
- Target temperature 33°C (range 32.5-33.5°C) in hypothermia group vs 36.5°C (range 36-37°C) in normothermia group
- Surface cooling methods used (forced-air blanket, optional circulating water mattress, optional IV cold saline)
- Cooling initiated after intubation and positioning, before likely time of surgical injury
- Rewarming began after final aneurysm clip secured
- Primary outcome: Glasgow Outcome Score of 1 (good outcome - mild or no disability) at 90 days
- No significant differences in primary outcome: 66% hypothermia vs 63% normothermia (OR 1.14, 95% CI 0.88-1.48, p=0.32)
- No significant differences in death rates at follow-up (6% both groups), ICU stay (mean 6 days both groups), total hospitalization (mean 16 days both groups)
- Postoperative bacteremia significantly higher in hypothermia group (5% vs 3%, p=0.05)
- No significant differences in rates of neurologic deficits, cardiovascular events, or coagulation problems
- Target temperatures achieved: 50% of hypothermia patients reached 34.5°C or lower within 1 hour before clipping
- Subgroup analysis showed possible benefit in males and patients with surgery 8-14 days after SAH, but these differences disappeared after adjustment for covariates
- Two planned interim analyses performed after 357 and 655 patients
- Nearly perfect follow-up rate (99.9%)
Study Design
- Study Type
- Prospective randomized controlled trial, partially blinded
- Randomization
- Yes
- Blinding
- All study personnel except anesthesiologists involved in intraoperative care were unaware of treatment assignments. Outcome assessors (neurologic examiners and neuropsychologists) were blinded to treatment assignment.
- Sample Size
- 1001
- Follow-up
- Approximately 90 days after surgery (median 88 days; 10th and 90th percentiles 72 and 113 days)
- Centers
- 30
- Countries
- United States, Canada, United Kingdom, New Zealand, Australia, Austria, Germany
Primary Outcome
Definition: Good outcome defined as Glasgow Outcome Score of 1 (minor or no disability) at approximately 90 days after surgery
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 314 of 501 patients (63%) | 329 of 499 patients (66%) | - (0.88-1.48) | 0.32 |
Limitations & Criticisms
- Patients cooled to only 33°C and only for intraoperative period - colder temperatures or longer cooling periods might yield different results
- Relatively slow rate of cooling may have resulted in inadequate protection during early stage of retractor placement or aneurysm exposure (though 50% achieved 34.5°C or lower within 1 hour before clipping)
- Esophageal rather than brain temperature controlled - potential discrepancies between measurements
- No control of postoperative care aspects
- Relatively crude outcome assessment (Glasgow Outcome Scale), though no differences found with other outcome measures either
- Study restricted to good-grade patients (WFNS I-III), limiting generalizability to poor-grade patients
- Obese patients (BMI >35) excluded, limiting applicability to this population
- Single-center pilot study findings may not generalize across all centers
- Subgroup findings (benefit in males and 8-14 day surgery group) were secondary analyses and must be interpreted cautiously
- Delayed surgery subgroup was small (93 patients) and imbalanced in baseline characteristics
- Neuropsychological examination results still being evaluated at time of publication
- Inability to completely blind outcome assessors to treatment (though they were independent of treatment teams)
- Two centers dropped during trial due to protocol compliance issues
Citation
N Engl J Med 2005;352:135-45