The first EEG does not always show an epileptiform abnormality in epilepsy and serial EEGs may all be normal.1 After the first EEG fails to show an epileptiform abnormality, the yield of epileptiform abnormalities can be increased if the EEG is repeated during sleep,1,2 but previous estimates of the yield vary due only to sleep.3–6 Hyperventilation and photic stimulation are also widely used as activating procedures on routine EEGs,7 although some authors have questioned their yield.8 Others have called for reevaluation of the usefulness of routine use of photic stimulation, even for generalized onset epilepsy.8–11
The aims of this study were to evaluate the contribution of activation procedures to the yield of epileptiform abnormalities and the predictors of finding epileptiform abnormalities in relation to activation procedures. To address these aims, we conducted a population-based study of newly diagnosed epilepsy, and assessed findings on serial EEGs performed from first diagnosis. No previous study of the yield of activation procedures has been conducted in a population-based setting on serial EEGs performed from onset. Use of this design provides important information for planning cost-effective diagnostic work-ups for patients with newly diagnosed epilepsy because it minimizes selection bias related either to epilepsy prognosis or to factors other than the presence of epilepsy. In contrast, studies restricted to cases from tertiary care centers may be biased by the inclusion of patients with more severe epilepsy, who have high seizure frequency and EEG abnormalities elicited either with or without activation.
The methods have been described in detail elsewhere.12 Briefly, data were obtained from the “Genetic Epidemiology of Seizure Disorders in Rochester” (GESDR) study,13–15 a population-based study using the resources of the Rochester Epidemiology Project (REP).16–20 The original study included all residents of Rochester, Minnesota with either a single unprovoked seizure or newly diagnosed epilepsy between 1935 and 1994. Our current analyses are restricted to patients diagnosed with newly diagnosed epilepsy between 1960 and 1994, and to EEGs recorded in 1960 or later. Only records with previous authorization for medical record review were included (more than 99% of those identified for the study).
Newly diagnosed epilepsy was defined as a history of two or more unprovoked seizures, separated by >24 hours and diagnosed by a physician. Subjects were classified as having newly diagnosed epilepsy if they were initially diagnosed with epilepsy while residing in Rochester from 1960 to 1994, or had a first unprovoked seizure from 1960 to 1994 and a recurrence at any time while residing locally during the follow-up period ending in 2008. For each resident with newly diagnosed epilepsy, we examined number of EEGs, age at diagnosis, sex, seizure type, etiology,21 presence of status epilepticus, and number of seizures per year of follow up. We excluded prolonged (longer than 2 hours), video- and intra-operative EEGs. The study was approved by the Mayo Clinic and Columbia University Institutional Review Boards.
For each EEG, trained nurses abstracted detailed information on date, activation procedures (sleep, photic stimulation, and hyperventilation), and activation-related findings. Two epileptologists (W.A.H., J.R.B.) reviewed the abstracted information to confirm it was correct. The EEG report did not indicate the duration of sleep, and whether sleep was spontaneous, induced with sleep deprivation, or only drug-induced was not abstracted. Similarly, no technical descriptions of hyperventilation and photic stimulation were provided. However, all EEGs and activation procedures were performed using methods that met or exceeded standard guidelines of the American Clinical Neurophysiology Society (www.acns.org/pdf/guidelines/Guideline-1), with a 21-channel 10-20 system.33 For almost the whole period of the study, the clinical neurophysiology laboratory of the Mayo Clinic was the only laboratory where EEGs were performed in the region. All EEG tracings were interpreted by board-certified clinical neurophysiologists. Information on antiepileptic drug use at the time of the EEG recording was not included in the EEG report.
Epileptiform abnormalities were defined by the presence of generalized epileptiform abnormalities (typical generalized 3 Hz spike-wave, atypical spike-wave, slow spike-wave, generalized epileptiform fast, hypsarrhythmia, and electrodecremental), focal epileptiform abnormalities (spike, spike-wave, sharp wave, periodic lateralized epileptiform discharges [PLEDS], multifocal, bilateral, independent, or synchronous), or epileptiform abnormality not determined whether generalized or focal.12
Activation-related epileptiform abnormalities were defined as those only due to the specific activation procedure regardless of the concomitant use and effect of other activation procedures on the same EEG. For each activation procedure, the EEG results were categorized as “procedure not done,” “procedure performed but no epileptiform/no additional epileptiform abnormality detected,” or “procedure performed and additional epileptiform abnormality detected.” An “additional epileptiform abnormality” was defined as any epileptiform abnormality not previously present, with or without an activation procedure.
Subjects younger than 1 year of age were excluded from the activation-related analyses because they are more likely to sleep spontaneously than older subjects and photic stimulation and hyperventilation are less feasible in infants.
Descriptive statistical analyses were performed to analyze demographic and clinical characteristics, using frequencies and percentages to summarize categorical variables, and medians and interquartile ranges for continuous variables. Statistical significance was determined using the χ2 statistic (or Fisher exact test) and Wilcoxon-rank sum test for continuous variables. The 0.05 level of significance and two-sided tests were used for all analyses. All analyses were conducted using SAS software, v 9.2., SAS Institute Inc, Cary, NC.
Cumulative Yield of Any Epileptiform Abnormality
We estimated the cumulative yield of epileptiform EEG abnormalities for the different activation procedures according to the number of EEGs using a life table approach for clustered observations, with clusters defined by individual subjects and observations defined by EEGs. Subjects were followed until the date an epileptiform abnormality was recorded, the date of last visit to an REP provider, the date of death, or the end of the study period (January 1, 2008), whichever came first.
We evaluated the yield of any epileptiform abnormality detected only through activation procedure(s) (i.e., sleep, photic stimulation, or hyperventilation) conducted at the time of the routine EEG. An abnormality was considered to be activation related if it occurred during an EEG at the time of a specific activation procedure and not in periods of the EEG without that procedure. Each activation procedure was considered, regardless of the use or effect of the other two on the same EEG. The analysis was performed by age group at diagnosis (1–19 years vs. ≥20 years) and by seizure type (excluding subjects classified as having both focal and generalized seizure types, because only two subjects were in this group).
The log-rank statistic was used to compare the estimate of cumulative proportion of epileptiform abnormality in newly diagnosed epilepsy among groups.
To determine the additional impact of activation procedures, we also considered the cumulative proportion with any epileptiform abnormality detected in awake EEGs where the epileptiform abnormality was not due to photic stimulation or hyperventilation. The analysis was repeated by seizure type and by syndrome, for subjects aged 1 to 19 years at diagnosis.
Cox Proportional Hazard Regression Analysis
A Cox proportional hazard regression model was used to evaluate the hazard of recording an activation-related epileptiform abnormality due only to any of the activation procedures with increasing time since diagnosis. In the model, adjusted for the number of seizures during the first 5 years of follow-up after diagnosis (categorized as none, one, two or more, or unknown), each activation procedure and age at EEG (dichotomized as 1–19 years of age and 20 years or older) were included as time-dependent covariates. Etiology was also included. A second model included the same variables and age at EEG was categorized as 1 to 9, 10 to 19, with 20 years or older as the referent.
To account for the clustering of EEGs within subjects, we used the marginal Cox model approach for clustered data, applying the Wei, Lin, and Weissfeld method.22
The proportional hazard assumption was tested graphically and by testing for interaction.
We studied 449 Rochester, Minnesota residents aged 1 year or older with newly diagnosed epilepsy from 1960 through 1994, who had at least one EEG. Among them, only 10 subjects (2.2%) had no activation procedure performed.
Cumulative Percentage of Subjects With an Epileptiform Abnormality Due to an Activation Procedure on Routine EEGs
Overall, activation by sleep yielded the greatest proportion of activation-related epileptiform abnormalities (Fig. 1A) followed by hyperventilation (Fig. 1B) and then photic stimulation (Fig. 1C). Within each activation procedure, the yield was greatest for individuals 1 to 19 years old.
The cumulative yield of epileptiform abnormalities due to sleep activation was similar in those with focal and generalized seizure types (18.5% and 18.0%, respectively, at the first EEG), whereas it was lower for those with unclassified seizures (3.2%) (Fig. 2A). For both hyperventilation and photic stimulation, the cumulative yield of epileptiform abnormalities was greater in those with generalized seizures (hyperventilation 20.5%, photic stimulation 16.0%) than in focal (hyperventilation 5.0%, photic stimulation 1.7%) or unclassified seizures (hyperventilation 0%, photic stimulation 1.6%, P < 0.0001; Figs. 2B and 2C).
Abnormalities Not Due to Activation Procedures
On the first EEG, the yield of ictal or interictal epileptiform abnormalities that were not due to activation, in awake EEGs, was 44% (see Figure, Supplemental Digital Content 1, http://links.lww.com/JCNP/A14). Among children aged 1 to 19 years, the cumulative yield of ictal or interictal epileptiform abnormalities that were not due to activation was similar in generalized and focal seizures (49% and 56%; see Figure, Supplemental Digital Content 2, http://links.lww.com/JCNP/A15), and greater in focal and generalized nongeneralized epilepsy syndromes compared with genetic generalized epilepsies (60% and 63% vs. 40%; see Figure, Supplemental Digital Content 3, http://links.lww.com/JCNP/A16).
Distribution of Epileptiform Abnormalities According to Activation Procedures
In subjects aged 1 year or older, 51.9% (N = 233) had no epileptiform abnormality on the first EEG, whereas 23.8% had an activation-related abnormality (due to sleep activation in 53.3%), and 24.3% had an abnormality that was not activation related (Table 1). Regarding abnormalities that were not activation related, sleep, photic stimulation, and hyperventilation might have been performed, but the epileptiform abnormality reported occurred only or also during wakefulness in the absence of these activating procedures.
Among 164 subjects who had a second EEG after a first EEG without any epileptiform abnormality, 17.1% (N = 28) had an activation-related epileptiform abnormality, mostly by sleep (64.3%). Only 8.5% had an epileptiform abnormality that was not due to activation (Table 1).
Among 82 subjects who had a third EEG after first and second EEGs without any epileptiform abnormality, 6.1% had an activation-related epileptiform abnormality and 7.3% had an epileptiform abnormality that was not due to activation (Table 1).
A complete distribution of activation procedures for each EEG is shown in Supplemental Digital Content 4 (see Table, http://links.lww.com/JCNP/A17).
Factors Associated With an Activation-Related Epileptiform Abnormality Across All EEGs
Adjusting for the number of seizures during the first 5 years of follow-up, the hazard of an epileptiform abnormality due to an activation procedure was increased for children aged 1 to 19 years (hazard ratio = 1.8, Table 2); sleep EEG also increased the hazard (hazard ratio = 2.7). The hazard was decreased for activation-related epileptiform abnormality in those with postnatal symptomatic etiology and those with unknown etiology compared with those with genetic etiology (Table 2).
Compared with the ≥20-year-old age group, in the 1 to 9 year olds, there was an increased hazard (hazard ratio = 1.8, P < 0.05) for epileptiform abnormality, whereas in 10 to 19 year age group, there was no statistically significant difference.
There was no multiplicative or additive interaction between age at EEG and each specific time-dependent activation procedure, implying that the combination of the activation procedure and young age was not associated with an additional increase in the hazard of epileptiform abnormalities.
This is the first population-based study of the yield of activation-related epileptiform abnormalities over multiple EEGs in people with newly diagnosed epilepsy. Previous studies on EEG activation procedures have focused mostly on selected patients with prevalent epilepsy. In these studies, the contribution of activation was not evaluated on multiple EEGs.
The literature is inconsistent regarding the ability of activation to trigger an epileptiform abnormality in different epilepsy syndromes and age groups.4,7,8,11,23–25 In our analysis of epileptiform abnormalities due solely to an activation procedure, the yield was highest for sleep and lower for hyperventilation and photic stimulation. In addition, the yield of activation was greater in children than in adults, across all three procedures.
Younger age at EEG recording and presence of sleep on EEG were significantly and independently associated with an increased likelihood of epileptiform abnormality, suggesting that the role of activation in detecting epileptiform abnormalities is greater in young people.
Sleep has been reported as an effective procedure to increase the yield of epileptiform activity after a normal EEG.26 Systematic studies of the yield of epileptiform abnormality in sleep have mostly considered sleep-deprived or unspecified EEGs,27 although sleep deprivation, spontaneous sleep, or drug-induced sleep are not always specified. Studies on children with incident seizure(s) reported a yield of 4.1% due only to sleep activation on the first EEG6 and 19.8% on a second EEG after sleep deprivation.1 In prevalent samples of patients with epilepsy, the yield on the first EEG varies from 11.4% to 22%,3–5 consistent with our findings (Table 3). A previous study found that older age at epilepsy onset was associated with fewer epileptiform abnormalities in both awake and asleep EEGs,3 consistent with our finding of higher yield in children than adults. Previous studies found that epileptiform abnormalities were more likely to be activated by sleep or sleep deprivation in patients with idiopathic generalized epilepsy than focal epilepsy,1,3,5 whereas we found that the yield did not differ between generalized and focal seizure types.
Only 5% of subjects in our study had an activation-related epileptiform abnormality due to photic stimulation on the first EEG. Although this was greatest in children and for generalized compared with focal seizures, the yield was low, especially when compared with sleep. Our findings are consistent with studies reporting a prevalence of photoparoxysmal response in epilepsy of 3% to 11%, higher in children than in adults and in generalized than in focal seizure types.11,28,29
The yield of epileptiform abnormalities is also low for hyperventilation compared with sleep. We found that the yield of activation-related epileptiform abnormalities accounted for 7.9% of epileptiform abnormalities on the first EEG, higher in children and in generalized seizure types. This result is similar to the existing literature, reporting an overall yield of less than 5%8 and of 6% to 11% in focal epilepsies7,30 (Table 3).
Detection of Activation in Serial EEGs
Our study evaluated whether the increased yield of any epileptiform abnormality after a first nonepileptiform EEG is due to activation procedures. We found that in subjects aged ≥1 year with epilepsy, who had a first EEG without an epileptiform abnormality, most activation-related epileptiform abnormalities on the second EEG were due to sleep.
Strengths and Limitations
Strengths of our study include its population-based design and the expertise of clinicians in and around Rochester. We evaluated each activation-related contribution to the yield of epileptiform abnormalities on sequential EEGs.
We analyzed each activation procedure independently of the others; however in many cases, multiple procedures were performed during the same EEG, as this is routine clinical practice. The exception is when photic stimulation and/or hyperventilation was contraindicated. Because of the way our data were collected (through retrospective medical record and EEG report review), it was not possible to determine interrater reliability of EEG findings and activation methods.
In keeping with usual clinical practice, the presence and type of activation-related EEG abnormalities in the first 6 months after diagnosis were taken into account in the classification of seizure type and epilepsy syndrome. This diagnostic issue creates a circularity of reasoning, especially regarding the yield of EEG abnormalities in primary generalized seizures and generalized epilepsy syndromes, where classification is greatly dependent on the identification of generalized epileptiform EEG abnormalities, and may have accounted in part for the greater yield we observed in generalized epilepsy (Fig. 2).
We were unable to address information on sleep deprivation, use of sleep induction drugs, and sleep structure was not abstracted and information on sleep duration was not available in the EEG reports. Also, it is unclear whether sleep deprivation before an EEG recording is the stimulus that lowers the threshold for epileptiform abnormality5,27,31 or whether sleep itself is the stimulus, and is facilitated by sleep deprivation.5,32
Among activation procedures, sleep is likely to be the most effective. After a first nonepileptiform EEG in which sleep was not recorded, a sleep EEG should be performed, in all age groups. Although clinically important when it occurs, the yield of photic stimulation and hyperventilation was low, especially in adults, even after serial studies.
Therefore for young patients, sleep, hyperventilation and photic stimulation should be performed if possible as the yield is greater than in adults. Given the low yield of hyperventilation and photic stimulation in adults, only a sleep EEG should be required, although photic stimulation and hyperventilation are easy to perform within the time allotted to a routine EEG. The value of ordering multiple EEGs over time, including activation procedures, should be considered carefully, to avoid potentially unnecessary testing in adults.
We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.
The authors thank Jane Emerson, RN, Melissa Petersen, RN, Diane Carlson, RN, Ann Van Oosten, and Thomas Bitz for assistance with data collection.
1. Carpay JA, de Weerd AW, Schimsheimer RJ, et al. The diagnostic yield of a second EEG
after partial sleep deprivation: a prospective study in children with newly diagnosed seizures. Epilepsia 1997;38:595–599.
2. Binnie CD, Stefan H. Modern electroencephalography: its role in epilepsy management. Clin Neurophysiol 1999;110:1671–1697.
3. Degen R. A study of the diagnostic value of waking and sleep EEGs after sleep deprivation in epileptic patients on anticonvulsive therapy. Electroencephalogr Clin Neurophysiol 1980;49:577–584.
4. Degen R, Degen HE. The diagnostic value of the sleep EEG
with and without sleep deprivation in patients with atypical absences. Epilepsia 1983;24:557–566.
5. Pratt KL, Mattson RH, Weikers NJ, Williams R. EEG
activation of epileptics following sleep deprivation: a prospective study of 114 cases. Electroencephalogr Clin Neurophysiol 1968;24:11–15.
6. Shinnar S, Kang H, Berg AT, Goldensohn ES, Hauser WA, Moshe SL. EEG
abnormalities in children with a first unprovoked seizure. Epilepsia 1994;35:471–476.
7. Miley CE, Forster FM. Activation of partial complex seizures by hyperventilation. Arch Neurol 1977;34:371–373.
8. Holmes MD, Dewaraja AS, Vanhatalo S. Does hyperventilation elicit epileptic seizures? Epilepsia 2004;45:618–620.
9. Baykan B, Matur Z, Gurses C, Aykutlu E, Gokyigit A.. Typical absence seizures triggered by photosensitivity. Epilepsia 2005;46:159–163.
10. King MA, Newton MR, Jackson GD, et al. Epileptology of the first-seizure presentation: a clinical, electroencephalographic, and magnetic resonance imaging study of 300 consecutive patients. Lancet 1998;352:1007–1011.
11. Wolf P, Goosses R. Relation of photosensitivity to epileptic syndromes. J Neurol Neurosurg Psychiatry 1986;49:1386–1391.
12. Baldin E, Hauser WA, Buchhalter JR, Hesdorffer DC, Ottman R. Yield of epileptiform electroencephalogram abnormalities in incident unprovoked seizures: a population-based
study. Epilepsia 2014;55:1389–1398.
13. Ottman R, Barker-Cummings C, Leibson CL, Vasoli VM, Hauser WA, Buchhalter JR. Validation of a brief screening instrument for the ascertainment of epilepsy. Epilepsia 2010;51:191–197.
14. Ottman R, Barker-Cummings C, Leibson CL, Vasoli VM, Hauser WA, Buchhalter JR. Accuracy of family history information on epilepsy and other seizure disorders. Neurology 2011;76:390–396.
15. Peljto AL, Barker-Cummings C, Vasoli VM, et al. Familial risk of epilepsy: a population-based
study. Brain 2014;137(Pt 3):795–805.
16. Hauser WA, Annegers JF, Kurland LT. Incidence of epilepsy and unprovoked seizures in Rochester, Minnesota: 1935-1984. Epilepsia 1993;34:453–468.
17. Hauser WA, Annegers JF, Rocca WA. Descriptive epidemiology
of epilepsy: contributions of population-based
studies from Rochester, Minnesota. Mayo Clin Proc 1996;71:576–586.
18. Melton LJ III. History of the Rochester Epidemiology
Project. Mayo Clin Proc 1996;71:266–274.
19. Rocca WA, Yawn BP, St Sauver JL, Grossardt BR, Melton LJ III. History of the Rochester Epidemiology
Project: half a century of medical records linkage in a US population. Mayo Clin Proc 2012;87:1202–1213.
20. St Sauver JL, Grossardt BR, Yawn BP, Melton LJ III, Rocca WA. Use of a medical records linkage system to enumerate a dynamic population over time: the Rochester epidemiology
project. Am J Epidemiol 2011;173:1059–1068.
21. Berg AT, Berkovic SF, Brodie MJ, et al. Revised terminology and concepts for organization of seizures and epilepsies: report of the ILAE Commission on Classification and Terminology, 2005-2009. Epilepsia 2010;51:676–685.
22. Wei L, Lin D, Weissfeld L. Regression-analysis of multivariate incomplete failure time data by modeling marginal distributions. J Am Stat Assoc 1989;84:1065–1073.
23. el-Ad B, Neufeld MY, Korczyn AD. Should sleep EEG
record always be performed after sleep deprivation? Electroencephalogr Clin Neurophysiol 1994;90:313–315.
24. Guaranha MS, Garzon E, Buchpiguel CA, Tazima S, Yacubian EM, Sakamoto AC. Hyperventilation revisited: physiological effects and efficacy on focal seizure activation in the era of video-EEG
monitoring. Epilepsia 2005;46:69–75.
25. Mendez OE, Brenner RP. Increasing the yield of EEG
. J Clin Neurophysiol 2006;23:282–293.
26. Gibbs E, Gibbs F. Diagnostic and localizing value of electroencephalographic studies in sleep. Res Publ Assoc Res Nerv Ment Dis 1947;26:366.
27. Mattson RH, Pratt KL, Calverley JR. Electroencephalograms of epileptics following sleep deprivation. Arch Neurol 1965;13:310–315.
28. Herrlin KM. EEG
with photic stimulation: a study of children with manifest or suspected epilepsy. Electroencephalogr Clin Neurophysiol 1954;6:573–589.
29. Roy AK, Pinheiro L, Rajesh SV. Prevalence of photosensitivity—an Indian experience. Neurol India 2003;51:241–243.
30. Gabor AJ, Marsan CA. Co-existence of focal and bilateral diffuse paroxysmal discharges in epileptics. Clinical-electrographic study. Epilepsia 1969;10:453–472.
31. Rowan AJ, Veldhuisen RJ, Nagelkerke NJ. Comparative evaluation of sleep deprivation and sedated sleep EEGs as diagnostic aids in epilepsy. Electroencephalogr Clin Neurophysiol 1982;54:357–364.
32. Degen R, Degen HE, Reker M. Sleep EEG
with or without sleep deprivation? Does sleep deprivation activate more epileptic activity in patients suffering from different types of epilepsy? Eur Neurol 1987;26:51–59.
33. Sinha SR, Sullivan L, Sabau D, et al. American clinical neurophysiology society guideline 1: minimum technical requirements for preforming clinical electroenephalography. J Clin Neurophysiol 2016;33:303–307.