COOL-MI: Cooling as an Adjunctive Therapy to Percutaneous Intervention in Patients With Acute Myocardial Infarction

Luis Gruberg, MD, FACC


November 25, 2003

Editorial Collaboration

Medscape &

Presenter: William W. O'Neill, MD, William Beaumont Hospitals (Royal Oak, Michigan)

Previous studies suggest that the proportion of ischemic risk zone that becomes necrotic following myocardial infarction (MI) is directly correlated with temperature. According to Hale and Kloner,[1] this fact suggests that immediate cooling immediately following infarct could yield potential therapeutic benefits. Using the rabbit infarct model, they consistently demonstrated in a series of studies that hypothermia can reduce myocardial infarct size. It has not been confirmed, however, whether the infarct size reduction noted in the rabbit model would also occur in the clinical setting and whether use of the technique would result in improved clinical outcomes. In addition, there are concerns that lowering the body temperature too much would result in an increase in patients' defibrillatory threshold.

Unlike previous surface cooling techniques, the Retrieve system (Radiant Medical; Redwood City, California) is an endovascular system that contains a closed-loop, heat-exchange catheter filled with normal saline that has the capability of cooling blood (to 33º Celsius) while circulating in the closed loop. This catheter is inserted percutaneously via the groin into the inferior vena cava, where cooling takes place by convection. The system has been approved by the US Food and Drug Administration for use during cardiac surgery.

The Cooling as an Adjunctive Therapy to Percutaneous Intervention in Patients With Acute Myocardial Infarction (COOL-MI) trial was conducted to test the safety and effectiveness of endovascular mild hypothermia using the Retrieve system as an adjunct to percutaneous coronary intervention (PCI) in the setting of acute MI. The primary hypothesis of the study is that the technique would have a beneficial effect in reducing final infarct size in this setting.


The COOL-MI trial randomized patients presenting with ST-segment elevation MI < 6 hours from symptom onset to either primary PCI or primary PCI with endovascular cooling. The cooling protocol was initiated in the cath lab and patients would undergo cooling to a target temperature of 33º for a period of 3 hours. As part of the cooling protocol, patients were administered buspirone (60 mg oral) and meperidine infusion at 25-30 mg/hr to control shivering. A forced air blanket was also used to warm the patient's chest and further suppress shivering. Rewarming was performed at 1º/hour. Patients with previous MI, cardiogenic shock, hypersensitivity to hypothermia, those taking buspirone or meperidine, or those with an inferior vena cava filter were excluded. The primary endpoint of the study was infarct size measured by Tc-99 sestamibi SPECT imaging and major adverse cardiac events (MACE) rate at 30-day follow-up. Secondary endpoints included:

  • ST-segment resolution at 90 and 180 minutes

  • Left ventricular ejection fraction

  • Creatine kinase and myoglobin (CK-MB) release


A total of 421 patients were initially included in the study; of those, 29 patients were deregistered and 35 patients were excluded due to catheter placement failure. The resulting final sample size included357 patients randomized to either hypothermia (n = 177) or to PCI only control(n = 180). The 2 groups were similar with regard to baseline clinical and procedural characteristics, and the majority of patients presented with inferior MI (Table).

Table. COOL-MI: Baseline Clinical and Procedural Characteristics
Characteristic Control(n = 180) Hypothermia
(n = 177)
Age (yrs) 59 ± 12 60 ± 12
Female (%) 26 24
Diabetes (%) 18 17
Hypertension (%) 39 47
Hyperlipidemia (%) 66 65
Smoker (%) 48 44
Prior MI (%) 12 10
Previous PCI (%) 13 12
Previous CABG (%) 2 3
Symptom onset to hospital (min) 238 237
Door to balloon (min) 92 110
Anterior wall MI (%) 44 42
Baseline TIMI flow 3 (%) 21 19
Final TIMI flow 3 (%) 95 90
Stenting (%) 92 84
Final blush score 3 (%) 33 29

CABG, coronary artery bypass graft; MI, myocardial infarction; PCI, percutaneous coronary intervention; TIMI, thrombolysis in myocardial infarction

Before 30-day follow-up, a total of 10 patients died and 22 patients did not undergo SPECT imaging. Final 30-day analysis was therefore, based on the 325 patients who had completed the study and had available SPECT data (Figure 1).

Figure 1. COOL-MI: study population.

Only 88% of patients achieved the target temperature prior to PCI; the average time required to cool the patient to < 35º C was 31 minutes. Overall, 94% of cooled subjects tolerated the decrease in temperature; the procedure was stopped in only 1 patient due to intense discomfort. At 30 days, the 2 groups had similar MACE rates and there was no difference in the incidence of death, reinfarction, or target vessel revascularization (Figure 2).

Figure 2. COOL-MI: 30-day MACE rates.

There were no statistically significant differences in adverse events between the 2 groups, although there was a trend toward more vascular bleeding, more cardiogenic shock, and higher TVR in the cooling group.

There was also no statistically significant difference in the rate of in-hospital adverse events such as pulmonary edema, shock, or arrhythmias between the groups, although the cooling arm showed a trend toward a slight increase in events (Figure 3). Changes in infarct size at 30 days, as assessed by SPECT, were also similar in the control and hypothermia-treated patients (13.8% vs 14.1%, respectively; P = .83).

Figure 3. COOL-MI: in-hospital adverse events.

Interestingly, when the data were analyzed for patients on the basis of infarct location, the infarct size of those with anterior MIs who had achieved hypothermia (< 35º C) (n = 16) was significantly smaller (9.3% of the left ventricle) compared with anterior MI patients who did not achieve the target temperature (21.9% of the left ventricle), and was also significantly smaller compared with control patients who presented with anterior MI (18.2% of the left ventricle) (P = .05). A trend toward lower CK-MB levels and higher left ventricular ejection fractions were also noted in anterior MI patients who achieved target cooling. No benefits were noted in hypothermia patients with inferior MIs.


Investigators drew the following conclusions from the COOL-MI data:

  1. Mild systemic hypothermia is safe and well tolerated during PCI therapy in acute MI.

  2. As administered in the present protocol, cooling did not result in a reduction of infarct size.

  3. An unexplained difference in efficacy of cooling between anterior and inferior MI exists.

  4. Patients with anterior MI who are cooled to ≤ 35° C appear to have a significant reduction in infarct size.


The concept of hypothermia exerting beneficial effects for the reduction of infarct size and the preservation of left ventricular function in acute MI patients is very attractive. The development of easy-to-use percutaneous devices that can make hypothermia possible has paved the way for full assessment of this technique. Unfortunately, the results of this trial did not fully support the initial hypothesis, although there seems to be a benefit in anterior wall MI patients who were able to achieve the target temperature. Interestingly, not all of the patients achieved the target temperature, and the procedure was well tolerated by patients with the aid of sedative drugs. Obviously, this initial study has raised interesting questions that will probably be addressed in future trials.

  1. Hale SL, Kloner RA. Myocardial hypothermia: a potential therapeutic technique for acute regional myocardial ischemia. J Cardiovasc Electrophysiol. 1999;10:405-413.