This essay will look at the prevalence of misdiagnosis in epilepsy in a variety of different settings, and will also attempt to identify the reasons why misdiagnosis is still highly prevalent. The negative effects of misdiagnosis will also be identified and some strategies to help reduce misdiagnosis rates will also be suggested.
Misdiagnosis: What is the risk and how does it arise?
Introduction
Epilepsy is defined as a recurrent tendency to have spontaneous, intermittent, abnormal electrical activity in the brain which manifests itself as seizures. Epilepsy is currently diagnosed after the occurrence of at least 2 unprovoked seizures. The diagnosis is heavily reliant on the clinical history provided by the patients and witnesses to their seizure events, however, tests such as EEGs can also help to provide useful information in helping to formulate a diagnosis. Approximately 1 in 131people in the UK currently suffer from epilepsy (456 000 people), and 1 person in 2000 is newly diagnosed with epilepsy every year (27 000 people).
One of the main challenges in epilepsy management is the high levels of misdiagnosis which still exist in practice. The current research shows a wide variation in the levels of misdiagnosis which exists, figures between 4.6% and 30% were encountered upon during the reading for this project.
This essay will aim to evaluate how common misdiagnosis is in epilepsy, and what are the reasons for why misdiagnosis may occur. I will also explain why misdiagnosis can be a problem for both patients and the health service as a whole. To conclude this report, I will also propose what steps clinicians can take to reduce the rates of misdiagnosis.
How common is misdiagnosis?
In my research for this project, one case kept cropping up in my reading and the case that has become synonymous with epilepsy misdiagnosis is the story of Dr Andrew Holton. Dr Holton was employed as a consultant Paediatrician at the Royal Infirmary Hospital in Leicester. During a period of employment of 11 years between 1990 and 2001, Dr Holton was responsible for diagnosing 1 948 children with epilepsy. During the latter part of the 1990s, the hospital in question was becoming inundated with both complaints and concerns being expressed by parents and also crucially fellow clinicians. Both groups expressed regret concerning the increasing number of cases of epilepsy being misdiagnosed by Dr Holton. The hospital trust was forced to act and suspended Dr Holton from his post in May 2001. During an internal review carried out by the trust, it was found that 619 cases out of the 1 948 diagnosed by Dr Holton (32%), the diagnosis of epilepsy was incorrect. A further 81 cases (4%) were still being investigated at the time of this report.
A misdiagnosis rate of 32% may be seen a very high, but I was surprised to read of some of the reactions of some doctors who had written into the BMJ to express concern about the punishment Dr Holton was receiving. Their reasoning was that the evidence from the Proceedings of the International League Against Epilepsy had also shown a similar rate of misdiagnosis in another study that was carried out across general paediatricians.
In fact, this concern expressed by other doctors was expressed in the final report to come out from the disciplinary proceedings against Dr Holton. The report commented 'that a misdiagnosis rate of approximately 1 in 3 may not be unusual', however they expressed serious concerns regarding the manner in which Dr Holton went about diagnosis cases of epilepsy in children. There were concerns regarding the ability of Dr Holton to interpret EEGs, and also the report had worryingly also found that Dr Holton had over treated approximately a third of his patients with Anti Epileptic Drugs (AEDs).
This case is very important as it helps to illustrate how common misdiagnosis may be in our NHS, but also it helps to show some of the ways misdiagnosis may occur. I was intrigued to see how common misdiagnosis is in both GP settings and also in specialist neurology centres.
A review article by Chowdhury et al set out to collate all the studies they could find regarding the incidence of misdiagnosis in practice. They analysed data from both adult and paediatric populations, and their literature searches only found 7 articles to review. As expected, a wide variation in misdiagnosis rates was seen.
The lowest rates of misdiagnosis found in the review article were from a Paediatric study carried out in the Netherlands. The study involved 412 children who had been diagnosed with epilepsy, being followed up over a period of two years. Only 19 of these children were found to have received an incorrect initial diagnosis, a rate of 4.6%
Contrast this with another study, carried out in Denmark also involving Paediatric patients. 223 children were reviewed who had been admitted to a tertiary epileptic centre. The diagnosis was disproved in 30% of cases.
The review doesn't seem to find any differences between misdiagnosis rates occurring in General practices, and epilepsy clinics. A study by Scheepars et al investigating diagnosis in 7 General practices, only found a misdiagnosis rate of 23%. A study by Smith et al carried out in a Liverpool neurology clinic, found that after retrospectively reviewing 324 patients, a misdiagnosis rate of 26% had existed in their sample.
However, we are all humans, and specialists still do mistakes. This was highlighted in a recent study aiming to find the efficacy of a new AED and it was found that of the 1 721 patients recruited to the trial, 44 (2.6%) of them had been incorrectly diagnosed with epilepsy by epilepsy specialists. This highlights the importance of continual review of an epilepsy diagnosis, to ensure no errors have occurred.
Why does misdiagnosis occur?
The main reasons why misdiagnosis occurs will now be explored in more detail.
The main reason why a diagnosis may be difficult is inadequacy in the patient's history. This maybe because the patient suffers from a loss of consciousness during the event, and therefore can't recollect what happened. Another reason is that there is a lack of a collateral history. It's inevitable that some seizures may occur without anyone being around to witness. However, even though a witness maybe present, there's no assurance that they may be able to assist you with a diagnosis. This is probably as a result of a witness who may have observed a seizure for the first time, and as a result their anxiety and fear at the time may cloud their ability to recount events fully later on. All these reasons are very important due to the reliance upon the history for the formation of a diagnosis.
The history may also be compromised in inadequacy of clinicians to obtain a full history from patients. A commentary by Ferrie et al, identified the fact that some clinicians are rushed into making a diagnosis, and fails to take the necessary steps to obtain a full complete history. They may not have time to track down witnesses to an epileptic seizure, or they may find it difficult to obtain a history from children. This is particularly the case when children suffer from sensory seizures, and it might be difficult to obtain a description of a visual aura for example.
Another reason highlighted by the Chowdhury review, is the effect that non-specialists in epilepsy have on increasing the misdiagnosis rate. A study carried out by Leach et al, set out to compare misdiagnosis rates between neurologists and non specialists, and the study found that neurologists had a lower misdiagnosis rate of 5.6% compared to 18.9% in the other cohort. It's as a consequence of results like this that have resulted in both NICE and SIGN to state in the epilepsy management guidelines, that the diagnosis of epilepsy in patients must now be confirmed by specialists. If specialists aren't available in the local setting, NICE states that referral to a tertiary epilepsy centre must be considered.
The list of differentials in epilepsy can be quite daunting, especially to an untrained clinician. Some of the features that are seen in an epileptic seizure may not be necessarily specific to epilepsy, and therefore it's important to acknowledge that you will see some overlapping clinical features between epilepsy and its differentials. This is further complicated by the absence of a validated diagnostic criterion to diagnose an epileptic seizure, even though the ILAE does produce a classification scheme of the different forms of seizures. As a result, as mentioned before, it's very important that specialist clinicians make the diagnosis, as they are best qualified to sort out the different differentials.
For example let's take the case of cardiac syncope and compare its features to epilepsy. Tongue biting is common in both, but one important fact to ascertain from both history and observation is where it occurs. Lateral tongue biting is 100% specific to generalised seizures according to a study carried out by the Cleveland clinic foundation. Also incontinence can be a feature in both, but is not specific to either. Studies have reported the incidence in epilepsy in 26-57% of events, while in it occurs in 44-60% of syncopal events. It's important to make the point that the SIGN guidelines have a useful table that aids clinicians in sorting out the different features obtained from the history across the different differentials.
In children, the NICE guidelines have listed 36 different conditions as a cause of a seizure. This again may be very daunting in inexperienced Paediatricians. It's again very important to sort out the common differentials in children from epilepsy, and again there's reliance upon the clinicians to take a complete history. This again highlights whether Paediatricians may be up to making a diagnosis. Currently, there was no requirement for Paediatricians to take specialist training in neurology let alone epilepsy. The question must arise then, how a Paediatrician be able to recognise a rare cause of a seizure? Ferrie argues in his commentary that this situation would not arise in diabetes or congenital heart diseases, so why should epilepsy be different.
The EEG is an important source of doubt in epilepsy diagnosis. One of the misconceptions about the EEG is that it's the gold standard tool in diagnosing epilepsy. This is incorrect, and the EEG lacks both specificity and sensitivity, therefore a normal EEG may not discount epilepsy while an abnormal EEG pattern may not be indicative of epilepsy either. The EEG must be analysed in association with the other forms of investigation (MRI, ECG, video telemetry etc) as well as the history.
It should be appreciated that approximately 10% of the general population exhibit non specific changes in their EEGs. This is important, as over interpretation of normal variant patterns in EEGs can be interpreted as being pathological of epilepsy, when in fact they are normal. A study carried out by Fattouch et al, concluded that interpretation on EEG findings constitute the main reasons for misdiagnosis between syncope and epilepsy in a sample of 62 patients. They found in their study that patterns such as slow delta waves and focal theta patterns were commonly confused as signs of epilepsy. This highlights that EEG abnormalities again are not uncommon in syncope patients.
Another study carried out by Benbadis et al, identified patients who had been incorrectly diagnosed with epilepsy, and they aimed to analyse what patterns in the EEGs of these patients caused the most difficulties in diagnosis. In the sample of 37 patients, the clinicians had identified 'temporal sharp waves' in 30 patients, along with frontal and generalised wave patterns in the others. The authors concluded that these patterns had been incorrectly identified and the patterns seen were 'benign simple fluctuations' of normal EEG patterns. These findings were echoed once again by a study by Smith et al, carried out in Liverpool, where the most common overread patterns were 'hypnagogic hypersynchrony and hyperventilation induced slowing'. This again highlights the intrinsic limitations of EEGs as they become exposed to subjective differences in interpretations between clinicians, particularly between non specialists and epilepsy specialists.
The prevalence of non specific EEG abnormalities was estimated to be around 3.5% in children in a study by Cavazutti et al. However the prevalence of these patterns increases in children with either behavioural or congenital abnormalities. 6% of children with ADHD showed epileptiform abnormalities in their EEGs, while 20-60% of children with autism were found to have EEG abnormalities, while 20% of children with congenital defects were found to have non-specific EEG abnormalities. The diagnosis of epilepsy in these children as a result would become further complicated due to these overlapping features, and as a result the EEG analysis would deserve specialist interpretation to reduce risk of misdiagnosis.
The abusal of EEGs is another factor to consider. A study by Stroink et al concluded that an EEG has no value whatsoever if there is no suspected epileptic attack from the history. The positive predictive value of the EEG in children where the diagnosis was unclear was only said to be 11.4%. Again, this reinforces the fact that basing a diagnosis of epilepsy on EEGs can lead to a large number of erroneous diagnoses.
This raises a final point I want to make, that there may be situations where clinicians may not have full access to the appropriate technology to make diagnosis. This may be more of a problem in developing countries, but the problem still applies in countries like the UK. Things like video telemetry can be very helpful in making diagnoses and they have been found to be more helpful than EEGs alone. It helps to eliminate some of the problems I identified in using witness accounts of seizure attacks, as a more qualified clinician may make a more accurate analysis and description of the attack.
Why is misdiagnosis a problem?
It's important to understand that not only does misdiagnosis carry physical effects on patients (i.e. side effects to unnecessary medication), but there are also psycho-social effects as well.
The patients and their families may become subject to unnecessary stigma from the community, but also restrictions on driving and employment are also placed upon patients. In children, educational expectations are lowered, thereby restricting their future career paths.
The AEDs themselves have been related to teratogenecity, but there have been accounts of severe life threatening hypersensitivity reactions. As well as causing congenital defects in the foetus, but AEDs also can cause an increased risk of spontaneous abortions.
Failure to diagnose epilepsy, can also delay treatment, thereby increasing the risk of seizure recurrence.
However, incorrectly giving a diagnosis of epilepsy can rarely cause a more severe diagnosis to be missed. This is very important for cardiac causes of seizures, as patients who suffer from syncope with underlying cardiovascular disease have a 30% mortality rate in the first year. Life threatening events such as prolonged QT interval forms of syncope must also be identified via the ECGs.
As well as decreasing public confidence in Doctors if misdiagnosis remains high, there's also a cost effect which must be analysed. A study by Juarez-Garcia et al estimated the cost to the NHS of £29 million, rising to £138 million. In times when cuts in public health funding are becoming necessary, this cost puts an extra burden on our NHS.
What can be done to prevent misdiagnosis?
Both NICE and SIGN have produced framework guidelines which help to aid all clinicians to reduce misdiagnosis. I've already identified a table produced by SIGN which identifies key points to look out for in the history which are relevant for the different clinicians (e.g. females with psychological history or stress are likely to suffer from Non Epileptic Attack Disorder). Both guidelines explicitly state that the final diagnosis must be confirmed by a specialist, and the research done by Leach et al echoes this point as specialists are less likely to make an incorrect diagnosis. The specialists themselves must carry out a significant proportion of their workload to epilepsy is advised by the SIGN guidelines.
The British Paediatric Neurological Association (BNPA) also conduct courses to help Paediatricians become more informed regarding epilepsy, and put them in a better state to make a diagnosis.
The review by Chowdhury et al, argue the case for 'acknowledging diagnostic uncertainty'. The reasoning for this is that it reduces the pressure placed upon clinicians to make a definitive diagnosis, and also reduces the need to make a rushed incorrect diagnosis. By reducing the pressure therefore we are decreasing the misdiagnosis rates. It's important however to communicate this with patients.
Conclusion - 300 words max
References - more than 20 please..
