Materials and Methods

Case selection—All cats evaluated for seizures at the Clinic of Small Animal Medicine at the Ludwig-Maximilians-Universität from 2000 through 2004 were eligible for inclusion.

Medical records review—Medical records were reviewed, and data extracted included signalment, age at the time of first seizure, seizure type (focal [primary or with secondary generalization] or generalized [tonicclonic or status epilepticus]), duration of seizure disorder, and results of clinical and neurologic examinations, hematologic and serum biochemical analyses, and ELISAs for blood-borne FeLV antigen and FIV antibody. Other sources of information included evaluations of images obtained via thoracic and abdominal radiography, abdominal and cardiac ultrasonography, and MRI or CT of the head; blood pressure measurements; and results of fundic examinations, CSF analysis, and histologic examination of the CNS.

Types of seizures were classified according to data in the medical records. A seizure was classified as focal when focal motor or autonomic signs with or without abnormal mentation were reported. When a focal seizure progressed to a generalized tonic-clonic seizure with loss of consciousness, it was classified as a primary focal seizure with secondary generalization. Primary generalized seizures were defined as loss of consciousness, motor activity involving the whole body, and lack of a seizure focus. Status epilepticus was defined as sustained seizure activity lasting for more than 15 minutes or serial seizures without complete recovery between seizures.

Seizures were further classified by etiology as reactive (metabolic or toxic), symptomatic (inflammatory, neoplastic, traumatic, or vascular), or idiopathic. Additionally, cardiac syncope was treated as a separate entity that causes episodes that closely mimic a seizure. Idiopathic seizures were defined as recurrent seizures of unknown origin accompanied by unremarkable findings from interictal neurologic examination and various other assessments (ie, tests of liver function; determination of thyroxine concentrations, FIV and FeLV infection status, concentrations of antibodies against Toxoplasma gondii, and blood pressure; fundic examinations; thoracic radiography; and abdominal ultrasonography). Classification of seizures as idiopathic was further supported whenever possible by unremarkable findings of analysis of CSF samples collected from the cerebellomedullary cistern (ie, protein concentration < 0.36 g/L and nucleated cell count < 5/μL; n = 13) and imaging of the head via CT or MRI (11). To determine the final outcome of cats that were discharged from the hospital, owners were contacted in June 2005.

Statistical analysis—Analyses were performed to evaluate differences in seizure type (focal, primary focal with secondary generalization, or primary generalized with or without status epilepticus), age at first seizure, and survival rate among cats according to seizure etiology (reactive, symptomatic, and idiopathic). The Fisher exact test was used to detect differences in seizure type. An overall F test and pairwise t tests were performed to evaluate differences in age at seizure onset. Survival times were displayed graphically as Kaplan-Meier curves and evaluated for differences among the 3 etiologic groupings by means of a log-rank test. To account for multiple testing, a log-rank test was performed in which the 3 groups were compared. Then, 3 log-rank tests were performed to compare pairs of groups. The level of significance was set at P ≤ 0.05.

Results

One hundred sixty-four cats with epileptic seizures were evaluated at the hospital from 2000 through 2004. Of all cats evaluated at the hospital during that 5-year period, 2.1% had seizure disorders. Sufficient data were available to classify seizures by etiology in 91 cats. Of those 91 cats, 53 (58%) were male (8 sexually intact and 45 castrated) and 38 (42%) were female (7 sexually intact and 31 spayed). Breeds of cats represented were domestic shorthair (n = 78), Maine Coon (6), Siamese (4), Persian (1), British Shorthair (1), and Devon Rex (1). Age at first seizure ranged from 1 week to 19 years (median, 5 years). This first seizure was observed between 1 day and 32 months prior to evaluation at the hospital; in 70% of the cats, it had occurred within a week of evaluation.

Physical examination—Clinical abnormalities were detected in 54 cats during physical examinations performed on admission to the hospital and daily thereafter. The most common abnormalities included signs of depression, hyperthermia related to seizure activity or inflammation or infection, anorexia, and heart murmur.

Interictal neurologic examination—Neurologic examinations were performed > 24 hours after the last seizure in 77 cats; results were abnormal in 43 (56%) cats and unremarkable in the remainder. The most common findings were abnormal postural reactions (n = 21), bilateral menace deficits (16), depressed mental status (12), anisocoria (8), ataxia (8), unilateral menace deficit (6), decreased facial sensation (5), nystagmus (5), tremors (5), decreased bilateral pupillary light reflexes (3), and compulsive circling (2).

Seizure characteristics—Seizures were classified as focal with or without secondary generalization in 52% of cats and as primary generalized with or without status epilepticus in 48% of cats. Etiology was classified as reactive in 22%, symptomatic in 50%, and idiopathic (n = 14) or presumptive idiopathic (9) in 25% of cats (Table 1). Cardiac syncope was diagnosed in the remaining 3% of cats.

Reactive seizures were diagnosed in 20 of 91 cats (22%) on the basis of abnormal results of laboratory testing or history of exposure to toxicants and unremarkable findings of interictal neurologic examinations. Associated metabolic disorders included hepatic encephalopathy (n = 4), severe uremia with end-stage renal disease (1), hypoglycemia attributable to insulin overdose (3), severe hyperthyroidism (serum thyroxine concentrations, 134 nmol/L and 67 nmol/L; reference range, 12 to 37 nmol/L) with severe systemic hypertension (2), polycythemia vera (2), severe anemia (PCV < 10%; 1), hypocalcemia (1), severe idiopathic hypertriglyceridemia (serum triglycerides concentration exceeding 10 times the upper limit of the reference range [reference range, 0.57 to 1.14 mmol/L; 1), and iatrogenic hypoadrenocorticism (1). There were 4 cats with intoxications; 2 had a recent (< 24 hours) history of permethrin (pyrethroid) spot-on application, and 1 had organophosphate exposure from an over-the-counter flea-and-tick collar and decreased serum cholinesterase activity (< 20% of the lower reference limit; reference range, 450 to 1.600 U/L). Intoxication was strongly suspected in a fourth cat on the basis of history and clinical signs but was not supported by results for cholinesterase activity. In 3 cats, events that initially mimicked generalized seizures were reclassified as cardiac syncopes because severe cardiac conduction disturbances were identified via ECG during (ventricular bigeminy with asystole up to 30 seconds) or shortly after the events (third-degree atrioventricular block). One cat had episodic collapse and the other cat had episodes of opisthotonus and collapse every 10 minutes, and ECGs were recorded between these episodes. Cardiomyopathy was diagnosed and intracranial disease was ruled out in 1 cat during postmortem examination.

Symptomatic seizures were diagnosed in 45 of 91 (50%) cats on the basis of results of CT (15) or MRI (1), CSF analysis (20), serologic testing for T gondii infection (5), measurement of blood pressure in combination with fundic and neurologic examination or postmortem histologic examination of CNS tissue (23). Inflammation of the CNS was the cause of the seizures in 44% (20/45) of the cats in this group. Etiologies included meningoencephalitis of unknown origin (n = 9), feline infectious peritonitis (5), toxoplasmosis (5), and feline infectious peritonitis and toxoplasmosis (1). Toxoplasmosis was suspected when there was a 4-fold increase in serum anti–T gondii antibody titers (n = 1) or when increased anti–T gondii serum antibody titers (IgM > 1:640 and IgG > 1:1,280) were associated with a mixed CSF pleocytosis (4). Diagnoses of feline infectious peritonitis were confirmed via postmortem examinations. Neoplasia was diagnosed in 38% (17/45) of the cats with symptomatic seizures. Eight cats had meningiomas, 1 had an astrocytoma, 1 had an oligodendroglioma, and 1 had osteosarcoma of the cranial bone. Metastasis of tumors to the brain was strongly suspected in 6 cats; however, this diagnosis was not confirmed via brain imaging because tumors had been detected elsewhere in the body and results of interictal neurologic examinations were suggestive of lateralized intracranial disease. Three of these cats had a thoracic mass, 1 had transitional cell carcinoma of the bladder, 1 had fibrosarcoma of the pinna, and 1 had an undefined abdominal tumor. Three other cats had a history of severe skull trauma. A vascular etiology was suspected in 3 cats with acute-onset seizures because of consistently high blood pressure (> 200 mm Hg) and extensive retinal hemorrhage in 2 of the cats and findings of lateralized deficits during neurologic examination and hypertensive angiopathy and encephalopathy during histologic examination of postmortem tissues from the third cat. No associated metabolic disorder was identified in these cats. Necropsy of 2 other cats with a recent onset of seizures but without status epilepticus revealed necrosis of the hippocampus and the piriform lobe associated with astrogliosis and sclerosis, respectively. There was no history of exposure to toxicants in these cats.¹³

Idiopathic seizures were diagnosed in 23 of 91 (25%) cats aged between 1 and 12 years (mean, 3.54 years; median, 3 years) on the basis of history of recurrent seizures and unremarkable results of physical and interictal neurologic examinations (23) and laboratory tests (including liver function tests [n = 20], CSF analysis [13], CT [7], or MRI [4]). Thirteen cats were evaluated at the hospital within 7 days after their first seizure occurred, whereas 10 cats had their first seizure 2 to 128 weeks before initial evaluation. Total duration of the seizure disorder ranged between 1 and 284 weeks (mean, 81 weeks; median, 47 weeks). In 9 of these cats, the diagnosis of idiopathic seizures remained presumptive because CSF analysis or brain imaging was not performed. These cats had recurrent seizures for up to 148 weeks (mean, 64 weeks; median, 54 weeks).

When owners were interviewed, 74% (17/23) of the cats with idiopathic seizures were still alive. Seven of those cats had 1 to 4 seizures/y for 72 to 284 weeks (mean, 174 weeks; median, 160 weeks). None had cluster seizures or status epilepticus. All had been treated with phenobarbital initially; however, owners of 4 cats did not follow through with the prescribed treatment plan because the cats refused the medication or were free roaming and returned home irregularly. Seizures subsided in 10 of the surviving cats. These were initially treated with phenobarbital after evaluation and, at the time of owner interviews, were still receiving the medication. Duration of episodes of seizures for those cats ranged from 1 to 108 weeks (mean, 30 weeks; median, 19 weeks).

Four cats with idiopathic seizures died or were euthanatized within 2 weeks to 32 months after their first hospital admission, 1 ran away 9 months after admission, and 1 was lost to follow-up. Those cats had recurrent seizures over a 2- to 196-week period preceding admission (mean, 56 weeks; median, 38 weeks). Physical and neurologic examinations, serum bile acid concentration and CSF analyses, and CT (n = 4), MRI (1), and histologic examination (1) had failed to reveal an underlying disorder. Three of these 4 cats subsequently died because of their seizure disorder.

Outcome of cats with seizure disorders—Owners were contacted between 6 and 54 months after initial admission to the hospital to obtain information on outcomes for the 91 cats with seizure disorders. For the 20 cats that had been evaluated for reactive seizures, cats lived between 1 day and 216 weeks (mean, 68 weeks; median, 35 weeks) from the initial onset of seizures, including 6 cats (30%) that lived to the time of owner interview. Eight others had died, and 6 were lost to follow-up.

The 45 cats that had been evaluated for symptomatic seizures lived between 1 day and 192 weeks (mean, 24 weeks; median, 3 weeks) after initial onset of seizures, including 4 cats that were still alive at the time of owner interview. Thirty-eight cats had died by the time owners were interviewed, and 3 were lost to follow-up.

The 23 cats that had been evaluated for idiopathic seizures lived between 2 and 284 weeks (mean, 128 weeks; median, 148 weeks) after initial onset of seizures. Seventeen of 23 cats of that group were still alive at the time the owner was interviewed and were included in the calculation, 4 had died (all with unremarkable results of CSF analysis and imaging of the head), and 2 were lost to follow-up.

Statistical analyses—Distributions of cats with different seizure etiologies (reactive, symptomatic, or idiopathic) among the various types of seizures (focal, primary focal with secondary generalization, or primary generalized with or without status epilepticus) were not significantly different (Fisher exact test, P = 0.83; Table 1). Although proportions of cats with generalized tonicclonic seizures appeared higher among those with idiopathic seizures versus those with other seizure etiologies, differences among the classifications were not significant. Likewise, although proportions of cats with generalized status epilepticus appeared higher in cats with reactive or symptomatic seizures versus those with idiopathic seizures, those differences were not significant.

Seizure etiology was significantly (F test, P = 0.005) associated with age at first seizure. Results of pairwise t tests suggested that cats with idiopathic seizures were significantly (P = 0.007) younger (mean, 3.5 years) when the first seizure occurred than cats with reactive seizures (mean, 8.2 years) and significantly (P < 0.001) younger than cats with symptomatic seizures (mean, 8.1 years). There was no significant (P = 0.99) difference in age at seizure onset between cats with symptomatic and reactive seizures (Figure 1).

View Full Size

Box and whisker plots of age of first onset of seizures according to etiologic classification in 88 cats evaluated for seizure disorders at a veterinary teaching hospital. Symptomatic seizures were defined as those with an inflammatory, neoplastic, traumatic, or vascular cause; reactive seizures were those with a metabolic or toxic cause; and all others for which a cause could not be determined were classified as idiopathic. Upper and lower limits of each box represent upper and lower quartiles, respectively. Horizontal line within the box represents the median value. Upper and lower whiskers represent the maximum and minimum values, respectively. Circle and asterisk represent outlying values. Mean age at first seizure is significantly (P ≤ 0.05) lower for cats with idiopathic seizures, compared with that for cats with symptomatic or reactive seizures.

Citation: Journal of the American Veterinary Medical Association 233, 10; 10.2460/javma.233.10.1591

• Download Figure
• Download figure as PowerPoint slide

Comparison of survival times among the cats via log-rank test revealed that seizure etiology had a significant (P < 0.001) effect on survival times. The 1-year survival rate for cats with reactive seizures (0.50) was significantly (P = 0.006) longer than that for cats with symptomatic seizures (0.16; Figure 2). Cats with idiopathic seizures had a significantly better 1-year survival rate (0.82) than did cats with reactive seizures (P = 0.02) or cats with symptomatic seizures (P < 0.001). Furthermore, cats with confirmed idiopathic seizures had a 1-year survival-rate of 0.74, whereas none of the cats with presumptive idiopathic seizures, in which CSF analysis and brain imaging was not performed, died during the study period.

View Full Size

Survival rate of 88 cats in which idiopathic (green line), reactive (blue line), and symptomatic (red line) seizures were diagnosed. Hatch marks indicate times at which cats were censored. The 1-year survival rate for cats with idiopathic seizures (0.82) was longer than that for cats with reactive seizures (0.50) and cats with symptomatic seizures (0.16).

Citation: Journal of the American Veterinary Medical Association 233, 10; 10.2460/javma.233.10.1591

• Download Figure
• Download figure as PowerPoint slide

A possible effect of status epilepticus on survival time was also examined. Results of a log-rank test that compared 1-year survival rates of cats with status epilepticus (0.00) and cats without status epilepticus (0.43) indicated that cats in which status epilepticus occurred had a significantly (P = 0.04) shorter survival time and status epilepticus occurred more commonly in cats with symptomatic or reactive seizures, compared with values for cats with idiopathic seizures (Table 1).

Discussion

In the retrospective study reported here, half (50%) of the cats with seizures had symptomatic epilepsy, whereas reactive seizures and idiopathic seizures were less common. Other researchers have suggested that intracranial disease is by far the most common cause of seizures in cats, accounting for up to 87% of seizures.^2,4 To those researchers, seizures of unknown origin represent symptomatic or probable symptomatic (cryptogenic) epilepsy on the basis of results of MRI or CSF analysis but without any histopathologic confirmation. In our study, symptomatic epilepsy was also the most common etiologic classification, but the proportion of cats in which symptomatic seizures were diagnosed was lower, more closely paralleling the findings of others who diagnosed seizures secondary to intracranial disease in 41%⁷ and 42%⁸ of cats with seizure disorders. It is possible that subtle intracranial disease was overlooked in our study because MRI of the head was not performed on each cat. Furthermore, seizures of unknown origin were classified as idiopathic seizures rather than probable symptomatic seizures.

Differences between results of our study and those of other researchers also exist in regard to percentages of cats with reactive seizures. Metabolic or toxic causes of seizures were identified in 22% of seizing cats that were admitted to the hospital. Others have reported considerably lower percentages.^2,8 Because MRI and CSF analyses were not performed for all cats, one could argue that some of the metabolic abnormalities detected in our study may not have caused the seizures and that additional intracranial disease could have been missed. Nevertheless, metabolic and toxicant-induced diseases (hepatic encephalopathy, insulin overdose, hypocalcemia, polycythemia vera, uremia, and permethrin and organophosphate intoxication) were evident in the cats of this report, and neurologic examination did not reveal signs of structural brain disease in any cat. Thus, it is unlikely that abnormal results of laboratory tests were incidental findings or consequences of status epilepticus. Other researchers included only cats with ≥ 2 seizures,⁴ thereby excluding the possibility of metabolic or toxicant-induced diseases, which may result in 1 seizure only.

The proportion of cats with reactive seizures could have been overestimated because associated metabolic disorders (eg, hypertriglyceridemia, end-stage renal disease, or hyperthyroidism) were included without regard to findings of MRI of the head. These metabolic disorders are uncommonly associated with the occurrence of seizures in cats.^8,14,15 Possible pathophysiologic mechanisms of seizure induction include cerebrovascular accident, hypertensive encephalopathy secondary to chronic systemic hypertension, serum hyperviscosity, and atherosclerosis with subsequent brain infarction or hemorrhage. In those conditions, seizures would not be true reactive seizures but, rather, symptomatic epilepsy secondary to cerebrovascular disease. Seizures occurred in 2 cats with severe hyperthyroidism, systemic hypertension, and left ventricular hypertrophy; therefore, cerebrovascular accidents may have occurred in those cats. On the other hand, results of neurologic examinations were unremarkable, and effects of excessive thyroid hormones include lowering the seizure threshold and affecting results of electroencephalography via their effects on γ-aminobutyric acid receptors, sodium channels, and glucose and oxygen metabolism.^16–19

Three cats of this study had cardiac conduction disturbances (third-degree heart block), and 1 cat had severe anemia and seizure-like syncopes. Nevertheless, initial recorded descriptions of these events closely mirrored descriptions of generalized seizures. Cardiac syncope is difficult to distinguish from seizures in humans.¹⁹ Cerebral hypoperfusion results in myoclonic jerks that resemble a short convulsive seizure. This clinical phenomenon has been named cardiac seizure and convulsive syncope in humans, although it is pathophysiologically distinct from a seizure.²⁰ Associated electroencephalograms are characterized by slowing of background rhythm, flattening, and high-amplitude delta-wave activity.²¹ Adding further to the challenge of distinguishing seizure from syncope, epileptic neural activity can provoke cardiac arrhythmias and ictal asystole.²² Thus, evaluation of simultaneous ECG and encephalographic recordings, a thorough clinical history, and cardiac assessment have been recommended to help make the distinction easier.²³

Necrosis of the hippocampus is a severe structural abnormality of the brain associated with seizures in cats in Switzerland and Italy.^13,24 There has been discussion whether this entity constitutes the cause or the consequence of severe seizures. Hippocampal neurons are sensitive to hypoxia, hypoglycemia, hyperglycemia following hypoglycemia, toxicants, and glutamate excitotoxicity.^25,26 In humans, ongoing loss of hippocampal neurons and astrocyte remodelling leading to temporal sclerosis is associated with initial precipitating injuries (eg, trauma, hypoxia, and prolonged initial seizure) early in life and perpetuated by ongoing seizure activity.^27,28 Genetic susceptibility of hippocampal neurons to excitotoxicity has been reported.²⁹ Because the cats in our study had no known exposure to toxicants and did not have generalized status epilepticus or severe metabolic changes, we chose to classify hippocampus necrosis as a structural abnormality of the brain related to epileptic seizures of symptomatic etiology.

In our study, 25% of cats were classified as having had idiopathic or presumptive idiopathic seizures. Although the existence of idiopathic epilepsy has been seriously questioned,^2,4 other researchers have reported idiopathic epilepsy in as many as 54% to 59% of seizing cats admitted to hospitals.^7,8,a

Differences among our results and those of other studies might relate to variation in referral population, inclusion criteria, group size, genetic composition of the population under investigation, duration of disease at time of evaluation, and components of the standard diagnostic evaluation. In our study, idiopathic seizures were diagnosed in cats for which no other cause of seizures could be determined. Mean age of onset of seizures was 3.54 years in those cats; this finding and others (longer survival times relative to other etiologic classifications and unremarkable results of interictal neurologic examination) were similar to characteristics of idiopathic epilepsy in dogs.^9–12

In dogs and humans, the term idiopathic epilepsy is used to describe epilepsy syndromes of presumed or defined genetic origin.^2,30 On the other hand, when an intracranial origin of seizures is strongly suspected despite unremarkable results of diagnostic tests, the term probable symptomatic or cryptogenic epilepsy is used.^2,4 There is no substantial evidence of genetic basis for epilepsy in cats, and we have been unable to link idiopathic seizures to any population of purebred cats; therefore, we avoided using the term idiopathic epilepsy. As suggested by others, it is possible that idiopathic seizures occurred secondary to nonsuppurative meningoencephalitis of unknown origin or cerebrovascular disease (thus representing probable symptomatic or cryptogenic epilepsy) or intoxication in a proportion of the cats in our study.^4,31,32 This hypothesis is supported by the finding that seizure disorders resolved after a short period in some cats.

In the study reported here, distributions of various types of seizures did not differ among cats with reactive, symptomatic, or idiopathic seizures. Generalized seizures, focal seizures, and focal seizures with secondary generalization occurred in cats regardless of etiologic classification. It had been postulated that focal seizures in dogs indicate structural brain disease in the contralateral hemisphere³³; however, this supposition was questioned by several researchers who also found high incidences of partial seizures in dogs with idiopathic and genetic epilepsy.^{10,11,34–37} In other studies^2,4,6 involving cats, generalized seizures were the most common type of seizure, including seizures in cats with brain tumors and hepatic encephalopathy, whereas in our study, focal seizures were more common.

Results of our study suggest that many cats with idiopathic seizures had a better prognosis than cats with intracranial disease or associated metabolic disease, yet life expectancy may be reduced in cats with idiopathic seizures versus healthy cats. Humans with idiopathic epilepsy live 2 years less, and those with symptomatic epilepsy 10 years less, than healthy humans.^38,39 Life expectancy is also shorter in dogs with epilepsy, compared with that of healthy dogs,^9,40 and dogs with status epilepticus have a poorer outcome and reduced survival time.^41,42 In our study in cats, status epilepticus was negatively associated with survival time and was more common in cats with symptomatic and reactive seizures. The short survival time of cats with symptomatic seizures was similar to results of another study² in which 6 of 7 cats with symptomatic seizures died within 3 months; however, a third study⁴ revealed long-term survival (range, 3 to 21 months) in 67% of cats with symptomatic epilepsy.

Investigators have postulated that non–feline infectious peritonitis viral encephalitis is a common cause of idiopathic seizures in cats.⁴ Their supposition was based on abnormal results of neurologic examinations or subtle CSF changes in their sample of cats. Other researchers have also diagnosed unclassified viral encephalitis in cats via postmortem evaluation.^31,32 We were unable to provide unequivocal support for this hypothesis in our study. Meningoencephalitis of unknown origin was diagnosed via CSF analysis and postmortem histologic examination in only 9 cats. Nonsuppurative meningoencephalitis has been associated with FIV, FeLV, parvoviral infections of the CNS, bartonellosis, cryptococcosis, and feline infectious peritonitis. Serologic tests for FIV and FeLV infections were routinely performed at our hospital and failed to provide any evidence for infection with these viruses as being a cause of the seizures. Nevertheless, we cannot rule out the possibility that some cats with idiopathic seizures suffered from a mild, self-limiting meningoencephalitis or vascular disease despite unremarkable results of neurologic examination or CSF analysis. Severe seizures occurred in a large proportion of cats with idiopathic seizures, continued with a lower frequency for several weeks or months, and then ceased after long-term treatment with phenobarbital, the preferred anticonvulsant in cats.^2,4,43,44

Our study had several limitations. We relied heavily on historical data from medical records, which may not have reflected the true variation in types of seizures among cats. Results of neurologic examinations may have been influenced by the degree of training of the individual performing those examinations. Most importantly, we had been unable to perform MRI of each cat with idiopathic seizures in the past. Computed tomography of the head is not as sensitive as MRI, and cerebrovascular disease and neoplasia other than meningioma may have been overlooked. Researchers that have used MRI to detect causes of seizures in cats postulate that cerebrovascular disease and nonsuppurative meningoencephalitis of unknown origin are common causes of seizures.¹¹ In the present study, we were unable to confirm or refute this hypothesis because results of MRI were not available for all cats. With the increasing availability of onsite high-field MRI units and the speed with which images can be acquired, more data regarding cats with suspected idiopathic seizures may become available to help to classify the cause of their seizures more precisely.