Previous studies have found that the devices work just fine in moving planes and ships. Jong Geun Yun of Dongkong College in Gwangju, South Korea, and colleagues found that when an ambulance was stopped, the devices they tested correctly analyzed heart rhythms generated on mannequins used in cardiac pulmonary resuscitation (CPR) training.

However, the devices "demonstrated significantly poor performance at all vibrations which can be routinely encountered during ambulance transport," the authors write in their report in the journal Resuscitation.

Based on the findings, "ambulance personnel should consider the possibility of misinterpretation...when the device is used while transporting a patient," they conclude.

The results suggest that there could be risks to using the devices in moving ambulances, said Dr. Richard Page, of the University of Wisconsin School of Medicine, Madison. "The bumps in the road could be misinterpreted" as a heart rhythm disturbance that needs to be shocked, Page said.

Also, in a moving ambulance, he said the devices could "undersense," meaning they could fail to recommend a shock when there should be one.

Still, most shocks to the heart don't happen in moving vehicles. "Defibrillation is rare in ambulances, but occasionally will happen," said Page, president of the Heart Rhythm Society.

"Nothing is perfect," he added, but defibrillators are very reliable. "Typically, correct sensing of the heart rhythm is not a problem since they are used at the scene."

Why the difference between planes and ambulances? Page - who did the planes study -- said it's likely because planes don't experience the kind of vibration that an ambulance would while moving.

The Korean research had three parts: a study of mannequins in a moving ambulance, a study of mannequins with simulated road vibrations and a study of six domesticated pigs with vibrations designed to simulate ambulance transport conditions.

At speeds varying from 20-100 kilometers per hour (12 to 62 miles per hour) and on unpaved roads at 10 kilometers per hour (6 miles per hour), the devices picked up rhythms in mannequins accurately.

During testing, researchers simulated cardiac arrest in the pigs, which have hearts resembling humans'. When they tested the devices while using vibrations similar to those found during the first part of the study, the devices misinterpreted many heart rhythms, more often in pigs than in mannequins.

Yun and colleagues evaluated two automated external defibrillators: Philips' HeartStart MRx, and CU ER 2 from CU Medical Systems Inc. in South Korea.

Ian Race, spokesman for HeartStart MRx manufacturer Philips Healthcare in Andover, Mass., said the company could not comment on the study without "reviewing the actual vibration profiles and individual event reports from the MRx used in the study."

However, he said that many factors can impact the quality of an ECG signal read by a defibrillator, including the degree of contact with the patient's skin and motion.

CU Medical Systems did not respond to a request for comment.

The Korean researchers suggested that dampening vibrations in ambulances and refining defibrillators might be in order. "These efforts are needed to strengthen the links of the chain of survival," Yun and colleagues write.

Or there's always treating the devices as you might a cell phone: Pulling over to the side of the road and testing again, said Page.

SOURCE: Resuscitation, online February 1, 2010.