Epilepia, 40(1 1):14841489, 1999 Lippincott Williams 8c Wilkins, Inc., Philadelphia 0 International League Against Epilepsy
Rapid Communication
Noninvasive Continuous Monitoring of Cerebral Oxygenation Periictally Using Near-Infrared Spectroscopy: A Preliminary Report P. David Adelson, Edwin Nemoto, *Mark Scheuer, *Michael Painter, *John Morgan, and Howard Yonas Departments of Neurosurgery and *Neurology, University o j Pittsburgh and Children’s Hospital qf Pittsburgh, Pittsburgh, Pennsylvania, U.S.A.
Summary: Purpose: To report on the use of near-infrared spectroscopy (NIRS) to examine the changes in cerebral oxygenation in the periictal period in patients with seizures. Methods: Cerebral hemoglobin oxygen availability was monitored continuously and noninvasively with NIRS in three patients (one in the pediatric intensive care unit (ICU) and two in epilepsy-monitoring units) in conjunction with continuous EEG monitoring. Ictal events were recorded and compared with the pre-, intra-, and postictal periods for cerebral oxygen availability, as defined by oxygenated hemoglobin (HbO,), deoxygenated hemoglobin (Hb), and the redox state of cytochrome oxidase (cytox). Results: Several important preliminary observations were made by using this technology. First, a preictal increase in cerebral oxygenation began between I and 2 h and >I0 h before the ictal event. Second, despite adequate perfusion, based on an observed increased HbO,, reduction in cytox indicates a per-
fusion-metabolism mismatch during seizure activity. Third, continued seizure activity and even isolated ictal events were associated with decreased cerebral oxygen availability. Fourth, differences in cerebral oxygen availability were noted between different types of seizures (e.g., electrographic seizures were accompanied by rapid reductions in HbO, and cerebral blood volume without reduction of cytox, whereas electroclinical seizures were characterized by marked increases in HbO, with or without reduction of cytox). Conclusions: In this preliminary report on the use of NIRS for patients with seizures, we believe that NIRS allows continuous and noninvasive monitoring of changes in cerebral oxygenation periictally, thereby permitting investigations into the pathophysiology of seizures and the exploration of the potential of cerebral oximetry as a tool for seizure localization. Key Words: Epilepsy-Surgery-Seizure localization-Nearinfrared spectroscopy.
Epilepsy remains a significant individual and societal problem. Advances in our understanding of the pathophysiology of seizures have resulted in improvements in our neuroimaging capabilities and localization of the region(s) of seizure onset noninvasively. At present, the two most commonly used imaging technologies for functional seizure localization are single photon emission computed tomography (SPECT) (1,2) and positron emission tomography (PET) (3,4). Although these imaging technologies have allowed great advancement in seizure localization and the understanding of epilepto-
genic zone metabolism, they provide only a single rneasure in time of cerebral blood flow (CBF) and metabolism. With the development of cerebral oximetry with nearinfrared spectroscopy (NIRS), the clinician has, for the first time, a method of continuously and noninvasively monitoring cerebral oxygen availability (5,6). Although NIRS has been shown to be clinically effective in monitoring cerebral oxygenation in various circumstances [e.g., during carotid endarterectomies ( 5 ) and after head injury (6)], its most useful application may be its ability to aid the further understanding of the pathophysiology of epileptic seizures; the pre-, intra-, and postictal perfusion-metabolism response. This preliminary report presents the cerebral physiologic results as measured by using NIRS in three patients who had a seizure(s) during monitoring.
Accepted July 12, 1999. Address correspondence and reprint requests to Dr. P. D. Adelson at Children’s Hospital of Pittsburgh, 3705 Fifth Ave., Pittsburgh, PA 15213, U.S.A. E-mail:
[email protected]
1484
1485
NIRS AND CEREBRAL OXIMETRY IN SEIZURES
MATERIALS AND METHODS Patient selection Informed consents were obtained from all patients after obtaining approval from the Institutional Review Board for the use of NIRS in studying the pathophysiology of neurologic disorders at our institutions. NIRS had been used as a routine part of an ongoing study of the physiologic changes occurring after severe traumatic brain injury (TBI) in children [Glasgow Coma Scores (GCS)] of 5 8 ) (7). Our initial observation was made in one patient who had generalized seizures a few days after his severe TBI. Prospectively, NIRS monitoring also was used on two patients with medically intractable seizures undergoing presurgical evaluation in the Epilepsy Monitoring Unit (EMU) for video EEG monitoring. In all instances, the optode was placed frontally on the forehead to monitor frontal metabolism and oxygenation. During the NIRS monitoring, a standard scalp EEG array was used to monitor one of the patients, whereas the other had implanted surface strip grid electrodes to monitor cerebral EEG activity. NIRS technique Cerebral oximetry data were obtained by NIRS with two instruments; the FDA-approved clinical device the INVOS3 l00A (Somanetics Corp., Troy, MI, U.S.A.), and the Investigational Device, the NIR0500 (Hamamatsu Photonics, Japan). The INVOS3 l00A cerebral oximeter uses a 2-wavelength light-emitting diode to measure the ratio of oxyhemoglobin (HbO,) to total hemoglobin levels (THb) in the brain, expressed as an absolute regional oxygen saturation (rSO,) index. The disposable sensor consists of a light source and two detectors located at 3 and 4 cm from the light source, allowing the correction for hemoglobin in the scalp and bone from the brain (5). The rSO, data from the INVOS3100A were captured on a 3.5-inch diskette. The NIR0500 has the transmitting and receiving optodes 4 cm apart, with a pathlength of 6 cm. The output is quantified as the change in micromolar (pM) concentration of HbO,, deoxyhemoglobin (Hb), and cytochrome oxidase (cytox). THb is the sum of HbO, and Hb, reflecting cerebral blood volume. The data were acquired on a 386 computer through a serial port by using the “Wedge” software (TAL Enterprises, Philadelphia, PA, U.S.A.). Theoretic basis of NIRS NIRS uses the ability of light in the near-infrared region of 600- to I ,000-nm wavelengths to penetrate tissue to depths of 8- 10 cm, and is used to monitor the absorption spectra of HbO, and Hb in cerebral blood vessels. As such, cerebral oximctry by NIRS monitors the balance between cerebral oxygen demand and supply (e.g., cerebral oxygen availability). If cerebral oxygen con-
sumption remains the same with increased oxygen supply, cerebral oxygen availability increases. If oxygen delivery decreases at constant demand, HbO, desaturation will be observed, and oxygen availability is reduced. The measurement of cerebral cytox, as can be done by using the NIR0500, is much more difficult than the measurement of HbO, and Hb because the cytox signal is one tenth that of the hemoglobin signal. Separation of the cytochrome signal from that for hemoglobin also requires the use of light at 4 wavelengths. The field of view of the sensor for the INVOS3 1OOA and the NIR0500 in the adult is estimated to be a hemispheric volume 4 cm in diameter, penetrating 4 cm into the head with -2 cm of that diameter skin and bone, and the other 2 cm, cerebral cortex.
RESULTSKASE REPORTS Patient 1 A 4-month-old boy had a GCS of 6 after suspected nonaccidental trauma. His EEG was being continuously monitored along with NIRS. At 40 h after admission, the INVOS3100A, with the optode on the right frontal region, revealed a progressive increase in cerebral oxygenation over 3 4 h, beginning at an rSO, value of 40 at 10:33 and increasing to a peak value of -70 at -13:00 (Fig. 1). Thereafter, rSO, fluctuated between 60 and 70 for 12 h until midnight, when the first seizure was observed, as indicated by the sharp increase in rSO, from 60 to 70, correlative with his electroclinical seizures. The seizures persistently recurred over the next 12 h until 12 noon the next day before they were finally controlled with multiple doses of antiepileptic medications (AEDs) including phenytoin (PHT; Dilantin), phenobarbital (PB), and lorazepam (LZP). His seizure frequency increased from about one per hour at onset to once every 2-3 min until broken at the end of 12 h. During this clustering of seizures, there was a progressive decline in rSO,. Patient 2 A 45-year-old woman had long-term medically intractable complex partial seizures for >20 years. She was admitted to the EMU for video-EEG monitoring along with NIRS by using the NIR05000. On the second day of her EEG monitoring at 20:00, HbO, and THb were noted to increase from -15 pM,whereas cytox was progressively reduced from 0 to -12 pM (Fig. 2) to increase to 50 when the seizure occurred at -21:42. Unfortunately, during the seizure, the patient inadvertently dislodged her EEG electrodes and the NIRS optodes from her head. These were replaced within 1.5 h, but immediately noticeable on reinstituting the data stream was that HbO,, Hb, and THb remained below baseline, even up to 4 h after the event; cytox returned to baseline at 3.5 h. Epilepsia, Vol. 40. No. 11, 1999
1486
P. D. ADELSON ET AL.
70
65
60 55 50 45
Patient 3 A 16-year-old girl had long-term medically intractable seizures since early after birth. Preoperative evaluation with video-EEG, MR, and PET could not localize nor lateralize the seizure focus. After discussion at the multidisciplinary epilepsy conference, she eventually underwent implantation of bilateral and interhemispheric surface electrodes in an attempt to Iateralize her epileptogenic zone. NlRS was continuously monitored during the
FIG. 1 . Cerebral oxygenation monitored by the lNVOS3100A (rS0,-index) in a 4-month-old boy with a Glasgow Coma Scale (GCS) score of 6 after suspected nonaccidental head injury. Spiking in latter half of tracing coincided with seizure activity verified by EEG recordings.
5 days before resumption of her typical seizures. On day 5 after implantation, she had one electrographic seizure; on day 6, she had three electrographic and one electroclinical seizures; and on day 7, one electroclinical seizure. All of her electrographic seizures were characterized by reductions in HbO,, Hb, and THb (Fig 3A); the electroclinical seizures differed in that HbO,, Hb, and THb increased during the seizure, indicating increased perfusion (Fig. 3B).
FIG. 2. Cerebral oxygenation monitored by the NIR0500 in a 45-year-old woman with intractable partial temporal lobe epilepsy for 20 years. Flat-line recording in the middle of the graph indicates period during which the optodes were torn from her head. HbO,, oxyhemoglobin; Hb, deoxyhemoglobin; THb, total hemoglobin; cytox, cytochrome oxidase redox state.
Epilqniu, V d . 40, N
