General Surgery among Long-Term Residents with and without Active Epilepsy at the Kork Epilepsy Centre – A Prospective Comparative One-Year Survey

Abstract

Background/Aims:

We prospectively assessed the frequency, type, severity and cause of treatment among the long-term residents of the Kork Epilepsy Centre in the year 2005.

Methods:

All long-term residents were exclusively referred to the general surgeon practicing on the campus. Patients were divided into 2 groups comprising 285 patients with active epilepsy and 53 controls who either never had epileptic seizures or have remained seizure free for at least 4 years.

Results:

The 1-year incidence of injuries was 54.2% among the active epilepsy group and 36.8% among controls (p = 0.0275). Several admissions due to differing reasons occurred in 16.8% of people with active epilepsy and in 7.5% of the controls (not significant). More than 2 admissions due to various reasons only occurred in patients with active epilepsy (range 3–9). Of all injuries, 41.3% were directly related to seizures. Among patients with active epilepsy, the 3 most frequent injuries were lacerations, bruises and fractures (29.6, 17.0 and 11.7%, respectively).

Conclusion:

According to this prospective study, active epilepsy turned out to be a significant risk factor for injuries under homogenous patient and observer conditions.

Introduction

Epilepsy belongs to the most frequent chronic diseases of the brain. The natural character of the disease implies an elevated risk of surgically relevant complications due to seizure-related accidents, though this is not yet proven [1]. Furthermore, the chronic use of antiepileptic drugs (AEDs), some of which are associated with a potential burden of influencing bone health and stability due to enzyme induction, suggests that patients having suffered from intractable epilepsy for years have a higher risk of suffering from complications requiring surgical treatment.

Since the elevated risk of injuries in epilepsy patients may warrant more caution and thus higher costs, both concerning necessary care at home and under working conditions, it is interesting to address the question of surgically relevant incidents. In the past, several studies have focused on this aspect. However, due to their design, the studies had methodological drawbacks: most of the previously published papers were based on multi-centre surveys from different countries, were retrospective studies, trusted in the feedback from various surgeons or surgical departments, did not homogenously assess the underlying epilepsy syndromes and clinically relevant comorbidities, partially relied on questionnaires of the patients and relatives themselves, or suffered from the lack of appropriate controls. Furthermore, most of the studies only assessed the immediately seizure-related injuries [2,3,4,5]. The frequency of seizure-related injuries per year in such studies ranged between 24 and 43% [3,4,6,7,8,9].

One section of the Kork Epilepsy Centre is the long-term resident department where more than 300 people live permanently. A detailed description of this section was published elsewhere [10]. Most of these residents suffer from severe additional handicaps like cognitive or central motor impairment beyond their epilepsy. However, some of them have never had seizures or have at least been seizure free for a long time. Since all of these patients are referred to the same surgeon (M.L.) who runs his private practice on the campus, we suggested that the prospective survey of visits to this surgeon would realistically reflect the incidence of surgically relevant assaults. Furthermore, the comparison between similarly handicapped people with active versus inactive or no epilepsy was performed in order to investigate additional and specific disease-related impact.

Patients and Methods

We prospectively collected all surgical consultations by long-term residents between January 1 and December 31, 2005.

In 2005, 338 long-term residents lived on the campus of the Epilepsy Centre Kork. All of these residents suffer from multiple physical and mental handicaps including epilepsy in most cases. They live together in apartments, attend our schools or work in our sheltered workshops if they are not retired. Educational and nursery staff is available at all times. The epilepsy treatment is supervised by 2 experienced full-time epileptologists (A.R. and G.Z.). In the case of accidents or injuries, the patients are exclusively referred to the general surgeon (M.L.) who runs his private practice on the campus. If he is not available, patients are first seen by the epileptologist on call. If he is not able to solve the medical problem, he will organize transportation to the surgical department of one of the other hospitals in the surrounding area. General medical care is provided by local general practitioners who are located near to the centre.

Including weekends and night services, the general surgeon was available on 243 days in 2005. He used a standardized assessment sheet which was filled out immediately after every visit. Since we wanted no input by other contributors to interfere with the homogenous data set obtained this way, we collected any additional severe injuries during 2005 separately (discussed later in the text), but did not include them into the prospective data collection set presented here. For incidence analyses, the figures per year were extrapolated from the data obtained by the general surgeon during his 243 days of attendance according to the formula X × 365/243.

Assessment

Details of the assessment sheet are shown in table 1. Traumatic symptoms comprised fractures, sprains, bruises, ligament ruptures, abrasive injuries, cuts, lacerations, nail dislocations and fractures.

Table 1.

Assessment sheet

1 Date and time of onset of the symptom leading to consultation

2 Date and hour of consultation

3 Surgical diagnosis according to the ICD-10; differentiation of traumatic or atraumatic symptoms

4 Correlation with epilepsy

4.1 direct

4.2 indirect

4.3 independent

5 Impact of epilepsy on surgical therapy

5.1 yes; if yes, give an explanation

5.2 no

6 Course

6.1 unknown

6.2 uncomplicated

6.3 complicated or prolonged due to epilepsy (option of explanation)

6.4 complicated or prolonged independent of epilepsy (option of explanation)

7 Outcome

7.1 treatment finished, no further requirements

7.2 patient lost to follow-up, no further consultation although this was recommended

7.3 no reconsultation necessary

7.4 delegation of treatment

We further documented the frequency and procedures of the surgical interventions.

We assumed a direct correlation between epilepsy and the symptom leading to the surgical consultation if the symptom resulted from an epileptic seizure. Examples for an indirect correlation are impaired coordination or ataxia due to high-dose antiepileptic drug therapy or fractures that occurred in spite of relatively harmless situations in patients being treated with enzyme-inducing AEDs for a long period and who suffer from osteoporosis. Further examples may be quoted in the following.

We suspected a direct impact of epilepsy on the therapeutic course if the necessary surgical treatment was impaired or prevented by a severe epilepsy- or drug-related psychiatric comorbidity, or in the case of extremely frequent or severe seizures.

The treatment course was scored as unknown if patients did not appear at follow-up consultations. An example of a complicated or prolonged course due to epilepsy would be an unsatisfactory fracture healing in the cases of very frequent or severe seizures.

Due to the homogenous distribution of physical and mental handicaps it appeared tempting to investigate the impact of epilepsy and its treatment on the risk of surgically relevant insults. For statistical analysis, the 338 residents were divided into 2 groups according to a seizure severity score we have used at our centre for several years. Since the degree of mental and physical handicaps is comparable in both groups, we preferred to compare these 2 groups methodologically in order to assess the impact of epilepsy and medical treatment on the probability of surgically relevant insults.

Scoring Categories

The scoring categories are shown in table 2. Group 1 comprised 53 patients (15.7%) without epilepsy or without active epilepsy (level 0–2). The severe psychiatric or neurological impairments of this group are similar to those of the other patients, and thus, they need to live in sheltered housing as well.

Table 2.

Classification of epilepsy severity

Level 0 = No epilepsy	− All GTCS
Level 1 = Inactive epilepsy, no therapy	− All seizures with loss or lowering of consciousness during wakefulness
Freedom of seizures for at least 4 years and no medication for at least 2 years.	− All seizures in patients with a history of serial seizures or status epilepticus
Level 2 = Epilepsy, seizure free with therapy
	Freedom of seizures for at least 4 years under stable AED treatment.
Level 3 = Epilepsy, low severity	Level 5 = Epilepsy, high severity
	Active epilepsy. High seizure frequency: >3 GTCS per month (average during the previous 12 months) or tonic, complex partial or simple partial seizures or GTCS or any other epileptic seizures with falls that occur more frequently than 5 times a month (average during the previous 12 months). Absence, atonic and myoclonic seizures of the upper body are classified as level 4 seizures if they occur as single seizures.
	Seizures with an elevated risk of injuries:
Active epilepsy. Seizures are mild according to their profile (frequency, symptoms, risk of serial seizures or status epilepticus, seizure-related safety hazards) so that they are a relevant risk factor only in well-defined and preventable situations (e.g., falls from relevant heights, swimming, driving, work at dangerous machines). Seizures usually do not require first aid or emergency drug application ictally or immediately postictally.
Examples:
− At least 3 bruises or distortions per year or
− 1 fracture or 1 first-degree head trauma during the preceding 5 years or
− Infrequent single tonic seizures exclusively during sleep	− 1 first-degree head trauma or an intracranial bleeding during the preceding 5 years
− Infrequent single weak simple motor or sensory seizures without propagation
− Life-threatening seizures that require immediate emergency aid by professional staff
− Infrequent and weak myoclonic jerks without loss of consciousness
− Patients who frequently suffer (cumulatively or for more than 12 h) from epilepsy–or drug-related loss of consciousness (postictal dreamy state, confusion, stupor). Patients with at least 1 status epilepticus of GTCS during the preceding 2 years
	− Infrequent single pure absence seizures without falls and within a predominant time frame (i.e. first 2 h after awakening)
− Infrequent short and weak tonic seizures in immobilized patients (who permanently use a wheelchair or permanently lie in bed)
Level 6 = Epilepsy, extremely severe
Epilepsy that would theoretically require a continuous hospital stay.
Level 4 = Epilepsy, medium severity
Active epilepsy, criteria for level 5 are not fulfilled. According to their profile, seizures require professional observation or first aid or the application of emergency drugs.	Examples:
– >15 GTCS per month
	Patients who have been seizure free for at least 4 years but are currently reducing AEDs.
− Frequent or longstanding (several days) loss or lowering of consciousness
Examples:	GTCS = Generalized tonic-clonic seizures.

Group 2 comprised all other patients (level 3 or worse; n = 285, 84.3%).

The primary study variable was the hypothesis that in group 2 patients (patients with active epilepsy), there was a significantly higher incidence of surgical admissions. Secondary variables addressed the hypothesis of significantly more admissions due to different reasons in group 2 patients and a significant contribution of the number of AEDs to the probability of traumatic lesions. Furthermore, we analysed whether the intake of enzyme-inducing AEDs had a significant influence on the frequency of fractures.

Statistical analysis was performed using SAS 9.1, based on the comparison of the ranking table of injuries per patient in both groups according to the Mantel-Haenszel (χ²) test for a linear trend on a 0.05 level [11].

Results

Among the total of 338 people, 160 were females (47.3%) and 178 males (52.7%). Age ranged between 11 and 93 years (mean 39.58 ± 14.80, median 39).

Two hundred and eighty-five patients (84.3%) had active epilepsy and therefore belonged to groups 3–6 according to the scoring system mentioned above. Fifty of them (14.8% of all 338 residents) belonged to group 3, 124 (36.7%) to group 4, 96 (28.4%) to group 5, and 15 (4.4%) to group 6, respectively. One hundred and thirty-five (47.4%) patients classified as group 3–6 were females. The age range was 11–82 years (mean 37.6 ± 13.62, median 38). All patients were treated with AEDs.

The control group (epilepsy classification group 0–2) consisted of 53 persons (15.7% of all residents). Eight persons (2.4% of the total of 338) had never had epileptic seizures, 15 (4.4%) had inactive epilepsy (group 1) and 30 (8.9%) had been seizure free for more than 4 years (group 2). Table 3 summarizes the data.

Table 3.

Distribution of epilepsy severity

n	Percentage of all inhabitants	Severity classification	Level
8	2.4	no epilepsy	0
15	4.4	inactive epilepsy	1
30	8.9	seizure-free under antiepileptic	2
		medication for at least 4 years
50	14.8	low severity	3
124	36.7	medium severity	4
96	28.4	high severity	5
15	4.4	extremely high severity	6
338	100

One hundred and forty-eight residents made at least 1 appointment with the surgeon (43.8%). Among patients with active epilepsy, 130 (45.6%) were admitted at least once, whereas 18 (34%) residents without active epilepsy made at least 1 appointment. Traumatic lesions were the cause of admission in 103 patients with active epilepsy (36.1%, 54.2% per year) and in 13 patients without active epilepsy (24.5%, 36.8% per year) (fig. 1).

Fig. 1.

Traumatic lesions (absolute percentage and annual rate).

One traumatic lesion occurred in 62 (21.8%) patients with active epilepsy and in 11 (20.8%) controls. Two traumatic lesions were documented in 21 patients (7.4%) with active epilepsy and in 2 (3.8%) controls. More than 3 different traumatic lesions only occurred in persons with active epilepsy: 3 in 10, 4 in 7, and 5 in 2 subjects. In addition, in 1 patient, 9 different traumatic lesions occurred. Data are shown in detail in figure 2.

Fig. 2.

Traumatic lesions per patient.

Statistical analysis addressed linear trends for an increased risk in group 1 (persons with active epilepsy) for traumatic lesions and revealed a statistically significant difference (p = 0.0275).

Having a closer look at the group of patients with active epilepsy, 50% of patients with less severe epilepsy (level 3), 46.7% of patients with medium severity (level 4), 30% of patients with severe epilepsy (level 5), and 53.3% of patients with extremely severe epilepsy had at least 1 traumatic lesion.

In the group with active epilepsy, the leading traumatic lesions were lacerations (29.6%), bruises (17.0%), fractures (11.7%), sprains (4.8%) and contused wounds (2.6%). In persons without active epilepsy, the 5 leading lesion types were bruises (15.4%), lacerations (11.5%), fractures (7.7%), burns (7.7%) and contused wounds, cuts and abrasive injuries (3.8% each).

In the case of active epilepsy, a direct impact of the underlying epilepsy was noted in 95 cases or 33.3%. The annual incidence was 50.1%. Among all appointments, the rate of directly related lesions was 41.3%. An indirect impact was scored in 33%. No relationship between epilepsy and lesions was assumed in 25.7%. Surgical therapy was affected by active epilepsy in 14 cases (6.1%). In 7 cases, the course was complicated or delayed due to the impact of the epileptic disorder. In an additional 4 cases (1.7%), complications occurred independently. Ten patients (7.7%) were on 1 AED at the time of the injury, the remaining patients were on 2 or more AEDs.

Self-evidently, no direct relationship was stated in patients without active epilepsy. However, due to long-term effects of AEDs, an indirect relationship was considered in 42.3% of all lesions. Twenty-five patients were on AED without enzyme-inducing activity (19.2%). The remaining 105 patients (80.8%) were treated with at least 1 AED with enzyme-inducing properties, for example with carbamazepine. The statistical analysis of the correlation between fractures and enzyme-inducing AEDs did not reveal a significant difference between the 2 patient groups (p = 0.4).

Since the general surgeon was not available during the whole year, we additionally collected the cases of severe injuries that happened during other times to avoid any relevant sampling errors. One ankle and 1 nose fracture occurred, both in patients with active epilepsy.

Discussion

The hitherto published papers that addressed the question of injuries in epilepsy have some methodological drawbacks. Questionnaires have the disadvantage of selection bias since affected persons will be more likely to answer actively than unaffected persons; multi-centre surveys naturally collect data from qualitatively heterogenous sites that may reflect various diagnostic and therapeutic attitudes and ratings; some of the most important surveys were retrospective, and therefore, were burdened by bias. Finally, most investigations raised the issue of seizure-related injuries only and did not include an appropriate control group in order to work out the specific impact of additional active epilepsy.

We suggested that the special conditions of our long-term department could help to overcome some of these drawbacks by performing a prospective survey within a very limited group of investigators who evidently have direct access to the patients living on the campus of the Kork Epilepsy Centre. Due to the complete prospective data acquisition, the frequency, severity and outcome of surgically relevant lesions could be assessed in a group of multi-handicapped persons, the majority of whom was suffering from difficult-to-treat epilepsy.

One could argue that our control group was not perfectly appropriate since level 2 patients who used to have active epilepsy in the past but have been seizure free for more than 4 years were included. Numerous reports have emphasized the possibility of long-term tolerability problems under treatment with enzyme-inducing drugs that might be relevant for injuries like fractures due to a progressive decline in bone mass [12,13,14,15]. However, in our study, enzyme-inducing drugs did not correlate with an increased fracture risk even in the active epilepsy group, nor did the number of AEDs in general, in line with another recent study [16]. Therefore, it is not probable that this factor might have played a major role in the overall results. Furthermore, the unequivocal effect of enzyme-inducing drugs on bone mass has recently been questioned [17]. Finally, one has to consider that other chronic drugs such as antidepressants and neuroleptics were frequently given in the whole cohort independent of an underlying epileptic disorder and could have equally interfered with the results [18]. Most probably, the cohort was too small for a considerable impact of any drug effect on 1 of the 2 patient groups.

Similarly to a large diary-based study that addressed morbidity in general [5], we found a higher incidence of admissions to our general surgeon in the group with active epilepsy (45.6 vs. 34%). One has to consider that admissions occurred not exclusively for traumatic lesions but also for other reasons such as the treatment of inflammatory diseases, abscesses or tumours. Still, most of the non-surgical disorders were treated by general practitioners and internists. Therefore, the higher percentage of annual medical treatments in the study of Van den Broek and co-workers [5] is not surprising (72% in the epilepsy group, 50% in the control group).

Looking at traumatic lesions only, which was certainly the most interesting aspect of this study, the annual rate was 54.2% among patients with active epilepsy and 36.8% among the controls, which is considerably more than in other studies that were based on questionnaires or telephone interviews and reported incidence rates of 21 versus 14 and 15% in an active epilepsy group [2,16]. We suggest that the continuous professional care and complete acquisition due to our referral system on the campus along with the fact that all of our residents suffer from various mental and physical impairments may have contributed to the considerably higher overall rate of injuries with a difference in favour of the controls that was comparable with the data of Beghi et al. [2]. In children with normal cognitive performance, epilepsy with almost 50% primary generalized epileptic syndromes did not increase the injury risk, whereas in the same study, children with an additional attention deficit hyperactivity disorder did have significantly more injuries than controls [18]. Clinically relevant comorbidity (as in both groups of our study) and sustained and difficult-to-treat epilepsy apparently elevate the risk of traumatic lesions [2,4,6,18,19,20,21,22,23,24], as confirmed by our data.

Injuries due to seizures have been reported to occur in about 15% of seizures [25]. In fact, in our group of 285 persons with active epilepsy, the rate was 11.7%, and thus, just 4% higher than in the control group consisting of persons with similarly pronounced physical and psychic handicaps. Most likely, studies based on reports of emergency departments are impaired by a severe selection bias and may therefore overestimate the real injury risk [19]. On the other hand, a higher injury risk in case of active epilepsy could just be confirmed in line with previous literature [2,7,20,26].

Some of the patients in our active epilepsy control group may have suffered from epilepsy not active enough in order to generate a more distinct difference according to surgically relevant insults as compared with controls. In the seizure frequency scoring system of the Mayo Clinic [19,27], with a score from 0 to 12, controls would have been scored 0–2, whereas the active group in our study would have been ranked according to scores between 3 and 12. However, we are fully aware that in the case of active epilepsy, around 40% of seizures may be overlooked [28,29,30]. In the prospective survey chosen, we intended to raise figures not interfered by any unsuspected seizures in the control group. Since no seizures or seizure-related injuries were observed among the controls, we conclude that our results are close to reality, under methodological restrictions we could not avoid. It is certainly possible that a comparable study among people not living in sheltered housing but suffering from equally active epilepsy may differ, as was the case in a recent survey among outpatients [16] that reported lower annual incidence rates of 15%. However, in our study, we addressed difficult-to-treat chronic epilepsy and not newly diagnosed patients.

A direct relationship between seizures and injuries was seen in a third of our patients with active epilepsy. This results in an annual incidence of 50.1%. In spite of the potential safety advantages of sheltered housing, this incidence is higher than that reported in the literature, varying between 24 and 43% [2,4,6,7,8,9].

Not only was the incidence of surgically relevant insults increased among patients with active epilepsy, but in patients with unfavourable epilepsy courses, surgical treatment may be delayed or impaired by ongoing epileptic seizures as it happened in 7 cases in this series.

In line with the literature, most injuries were minor and transient [2,3,4,6,19,25,31,32]. In most instances, it was not necessary to refer patients to other hospitals, and surgical treatment could be finished on the campus.

Fortunately, sudden unexplained death of epileptics or other epilepsy-related fatal incidences such as drowning did not occur. Drowning has repeatedly been reported as the most common cause of fatal injury due to seizures, especially in children and young adulthood [33,34,35,36]. Of course, this was partly due to an accidental sampling artefact since both sudden unexplained death of epileptics and drowning occasionally occurred in the history of our long-term resident institution. Drowning may also be less frequent since in every case of active epilepsy, our regulations recommend that experienced staff have to accompany a person at risk while bathing or swimming.

In conclusion, this survey using a hitherto unique comparative study design supported the evidence that epilepsy with medically insufficiently controlled seizures is a risk factor for increased admissions to general surgery.