Abstract
Summary: Objective: To demonstrate that seizure frequency in patients undergoing video EEG telemetry does not correlate with atmospheric pressure (AP) changes.
Method: Historical automated AP data from weather stations in the Seattle Metropolitan area were correlated to seizure frequency and type in consecutive patients undergoing video EEG telemetry at our institution from April 2005–April 2006. Daily maximum, minimum and range of atmospheric pressures were correlated to daily number of events (seizures, pseudoseizures, unknown) per patient. Alternatively, whether or not events occurred during a change of greater or less than 5.5 mBar per day were used to evaluate odds ratios of events occurring.
Results: Of 191 patients, 96 were diagnosed with epilepsy, 60 with pseudoseizures, and 40 had different diagnoses. A total of 159 seizures, 59 pseudoseizures, and 40 unknown events occurred. No correlation between daily mean, maximum or minimum pressure change with seizure or other event frequency was seen. With increased daily AP range an increase in daily seizure per known seizure patient occurred (P = 0.03). Patients with known epilepsy showed an OR of a seizure occurring of 2.80 (95% CI 1.22–6.42 P = 0.02) if pressure changed more than 5.5 mBar that day.
Conclusions: Surprisingly, in patients with known epilepsy, increased seizure frequency occurred with changes in barometric pressure, particularly over 5.5 mBar range per day. Speculative mechanisms of AP change on seizure susceptibility are discussed.
BACKGROUND
One of the great riddles in epilepsy is why some patients develop seizures after hyperventilation (HV). Innate HV responses occur in various settings including high altitudes, diminished partial pressures of O2, elevations in CO2, and increased metabolic demand (West, 2004). Atmospheric pressures (AP) directly influence partial pressures of O2 and furthermore, pressure alterations are associated with human seizures, though mainly triggered within artificial environments of high-pressure hyperbaric chambers (Boldrey and Millichap, 1966; Hampson and Atik, 2003; Yildiz, 2004). Despite this observation, few objective studies specifically address if differing atmospheric pressures alter seizure frequencies. The studies that have, utilized self-reported diaries or ER presentations (Asensi et al., 1977).
We hypothesized daily atmospheric pressures would not alter seizure frequency in patients undergoing video EEG telemetry (VEEG) at our institution. We set out to test this hypothesis using public access Seattle-specific atmospheric data and our VEEG telemetry event frequency data.
METHOD
Consecutive hourly atmospheric pressure data spanning April 2005–April 2006 from automated National Oceanographic and Atmospheric Administration stations (NOAA) were downloaded into a statistical package (SPSS V 11.0.1, Chicago, IL). Data from two stations were used: the first, a sea-level buoy in Seattle’s Elliot Bay approximately a half-mile west and three-hundred feet below our telemetry unit (47°36′18′ N 122°20′18′ W) and the second a land- based station three miles nor-nor-west and 290 feet below our telemetry unit (47°39′44″ N 122°26′09″ W) (National data buoy center, 2006). Data were parsed to include daily high, low, mean, and range of atmospheric pressure recordings for the corresponding day two of VEEG. The data were not corrected for altitude as we assumed the correction would be a constant and of no meaning in correlative analysis.
Seizure frequency and other data from day two, by our definition Tuesday, of consecutive patients undergoing VEEG were collected over a year. Day two of telemetry was chosen so as to maximize patient numbers and minimize effects of residual antiepileptic drugs (AED). AED taper regimens are patient-dependent in our unit, historically 95% of our patients, including those with suspected pseudoseizures, have their anticonvulsant medications stopped at admission day one. The average length of stay in our VEEG unit is four days. By choosing day two various potential confounders can be essentially randomized: these include alternative drug taper regimens, and minimization of patients preparing for hospital discharge and as such only undergoing a half day of VEEG. The latter consideration becomes particularly important when including our control pseudoseizure group, who do not stay as long as our epilepsy patients. Collected values included the number of patients monitored, the number of seizures each patient had, and the classification of seizure (epileptic, psychogenic, or unknown). Events were classified by multidisciplinary conference review; if there was any uncertainty as to an event it was classified as an unknown event.
Pearson’s rank sum correlations of mean, maximum, minimum, and atmospheric pressure ranges with event frequency were performed (Table 1) as were correlations with event frequency per group (i.e., number of seizures per patient with epilepsy on that day) (Figure 1). In order to minimize effects of multiple sampling errors or outlying effects from patients with extremely frequent seizure, odds ratios (OR) were also calculated. Either the median pressure range (5.5 mBar) or one standard deviation above median (9.3 mBar) served as dichotomization points i.e., daily range ≤5.5 mBar, or >5.6 mBar. This was used to then classify if patient with known epilepsy, pseudoseizures, or unknown spells had an event that day. The p-values were conservatively calculated using two-tailed Fisher’s exact. Additional study of daily atmospheric pressure trend (rising greater than 4 mBar per day, declining by more than 4 mBar per day or no obvious trend in rising or falling pressure) were similarly evaluated. Significance was determined if p < 0.05. Our institutional review board approved this study. […]
Figure 1. Scatterplot of daily seizures per epilepsy patient plotted against range of barometric pressure in 24 period.
TBI Rehabilitation 