PARTICIPATION of anesthesiologists in murder trials is probably rare. I became involved in a case of an anesthesiologist prosecuted for murder by application of succinylcholine. The case investigated whether electrocardiogram strip chart recordings, showing ventricular fibrillation and presented by the charged physician as being recorded from his wife during cardiopulmonary resuscitation (CPR), were authentic or faked with a simulator.
A 35-yr-old woman was found dead by a mobile physician and paramedic rescue team at the bottom of a stairway next to a vacuum cleaner. Prolonged CPR was unsuccessful. Eight minutes earlier, the team had been dispatched to the scene, a couple's home, via telephone alert by the woman's husband, a trained anesthesiologist. He reportedly had found his wife unconscious and looking pale, with a slow and weak pulse that eventually faded. Before the alert, he himself had administered CPR, including tracheal intubation and bag ventilation using his emergency equipment. Furthermore, he presented several electrocardiogram strips with date and time annotations showing consecutive sinus bradycardia (exactly 30 beats per min), asystole, and ventricular fibrillation (fig. 1), recorded before the team's arrival using his portable electrocardiogram monitor and printer, and he suggested that his wife's blood pressure problems possibly evoked her accident.
After the woman was pronounced dead, police officers were called and performed a workup and seized materials, including the electrocardiographic recordings. Of note, one policeman noted that the dead woman's husband might have tried to conceal one of the electrocardiogram strips.
On forensic autopsy, some pulmonary edema, a scalp laceration, and findings consistent with intubation and CPR were seen, but no pathology indicating inflicted external force or explaining the woman's death. Accordingly, after taking specimens, the woman's corpse was released for cremation. On later review, however, the possibility of murder rather than an accident was considered because of a suspected discrepancy between the initial sinus bradycardia and the stated but unsuccessful resuscitation measures, a succinylcholine vial considered missing from the anesthesiologist's emergency case, and testimony regarding the couple's seriously disturbed relationship. The anesthesiologist was charged with murder by using succinylcholine.
However, no succinylcholine or its degradation products were found in specimens by using mass spectrometry. Subsequently, the court sought my testimony on whether electrocardiographic recordings could have been generated using an electrocardiogram simulator (Heartsim 2000; Laerdal, Stavanger, Norway) available to the accused at his hospital.
In response, I recorded simulator rhythms on a similar electrocardiogram recorder with all lead permutations in the recorder's diagnostic and monitor modes and compared them with electrocardiographic strips seized from the scene. Furthermore, I asked the simulator's manufacturer to provide simulated recordings and to indicate how they had been originally obtained. Of note, the tracing showing ventricular fibrillation secured on the scene bore a striking resemblance with one recording from the simulator. In fact, both tracings showing ventricular fibrillation were essentially superimposable when using imaging software (fig. 1). Furthermore, the fibrillation pattern appeared to be repetitive after 4 s. Accordingly, I testified that at least part of the electrocardiographic recordings secured on the scene, previously presumed to have been recorded from the accused's wife during CPR, was identical with the unique ventricular fibrillation pattern stored in and displayed by this particular simulator. Ruling out random coincidence, this implied that electrocardiogram tracings seized on the scene had not been recorded from the woman found dead.
Subsequently, the anesthesiologist admitted to having killed his wife, without elaborating on the circumstances for legal reasons, and was sentenced to prolonged imprisonment.
This case is interesting from a medical, educational, and legal perspective. Certainly, intravenous or intramuscular succinylcholine can evoke prolonged respiratory muscle paralysis 1,2and asphyxia within minutes. However, because it is rapidly cleaved in blood and tissues and breakdown continues after death, succinylcholine can hardly be detected in specimens by routine forensic mass spectrometry. 3,4Accordingly, succinylcholine was not detected in the remains of five victims of a confessed serial killer using intramuscular suxamethonium, but was detected in a syringe buried nearby. 5With no proof of succinylcholine administration, the prosecutor's line of evidence rested on the investigation of whether electrocardiographic recordings, previously stated by the accused to have been recorded personally from his wife during CPR, were faked, thus implying a criminal act.
Ventricular fibrillation is recognized by erratic activity around the electrocardiogram's baseline with no regular pattern, but possible periodicity has not been studied in humans. However, fibrillated animal hearts show little evidence of relevant periodicity. 6,7Thus, any repetitive pattern during an epoch of apparent fibrillation suggests artifact rather than authentic cardiac activity.
Electrocardiogram simulators have stored in their memory epochs of rhythms once recorded from “real” patients. It is less appreciated, however, that these epochs are displayed as endless loops and, hence, display periodicity not present in original recordings. With respect to ventricular fibrillation, this simulator displays an identical fibrillation pattern every 4 s (fig. 1). Accordingly, the same periodicity in the secured tracing suggested an origin from a simulator rather than authentic fibrillation. Furthermore, fibrillation tracings were virtually superimposable. In fact, I considered the chance that an epoch of fibrillation had shown both the same periodicity and the same pattern morphology as a specific simulator's fibrillation tracing as zero, suggesting that the electrocardiogram (including date and time annotations) was faked. Of importance, filters used in the electrocardiogram recorder's diagnostic and monitor modes substantially changed the fibrillation pattern, and without painstaking comparisons, using all leads and filters, the identity of the tracings easily could have gone unnoticed.
In summary, this scurrilous case shows that not all apparently medical evidence can be taken to be real and that criminal action may involve the use of faked electrocardiograms using medical simulators.