Q J Med 2004; 97: 337-341
QJM vol. 97 no. 6 © Association of Physicians 2004; all rights reserved.
Inappropriate requests for serum anti-epileptic drug levels in hospital practice
From the 1Epilepsy Unit, Department of Neurology, University Hospital of Wales, Cardiff, 2Therapeutics and Toxicology Centre, Llandough Hospital, Cardiff, and 3Walton Centre NHS Trust, Liverpool, UK
Received 9 December 2003 and in revised form 11 March 2004
 |
Summary |
|---|
 Top Summary IntroductionMethodsResultsDiscussionReferences |
|---|
Background: Serum anti-epileptic drug (AED) levels are indicated to assess AED adherence or toxicity, and are applicable to only a few AEDs. Expert consensus views on the clinical role of serum AED levels are summarized in the evidence-based guidelines published by the Scottish Intercollegiate Guidelines Network.
Aim: To examine local compliance with these guidelines.
Design: Retrospective case-note audit.
Methods: We included all serum AED level measurements requested from our hospital over two months. Our audit standards were first, that serum AED levels should be requested only for suspicion of poor AED adherence or toxicity (‘indication-compliant’), and secondly, for ‘full compliance’, that ‘indication-compliant’ requests should be made only for AEDs with established dose-response and dose-toxicity relationships (phenytoin, carbamazepine, phenobarbitone).
Results: There were 114 measurements in 102 patients. Serum AED level requests were for phenytoin (n = 50), valproate (n = 27), carbamazepine (n = 22), lamotrigine (n = 8), phenobarbitone (n = 7), and were made by physicians (n = 46), paediatricians (n = 30), neurologists (n = 15), neurosurgeons (n = 14), psychiatrists (n = 7), and intensivists (n = 2). AED toxicity was queried in 29 requests (25%), and adherence in 10 (9%); thus 34% of requests were ‘indication-compliant’. However, 16 of these were for valproate or lamotrigine; thus only 23 requests (20%) were ‘fully compliant’. Clinical management changed in only 17 of the 47 patients whose levels fell outside target ranges, and only two of these followed indication-compliant AED measurement.
Discussion: The audit identified a failure locally to comply with standard evidence-based guidelines. If, as is likely, this reflects practice elsewhere in the UK, there are potentially major clinical management and resource implications.
|
Introduction |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Serum anti-epileptic drug (AED) level monitoring has traditionally been used in epilepsy management, owing to wide variations in pharmacokinetic handling of AEDs between individuals. It is particularly useful in guiding changes to phenytoin therapy, because of phenytoin's zero-order kinetics and narrow therapeutic index. Medical and surgical specialists currently request AED levels for a variety of reasons,2–4 but the tests are often unnecessary or inappropriate, risking erroneous interpretation and clinical management.5 For example, seizure control may be perfect despite AED levels below the target range, or there may be no clinical toxicity despite serum concentrations above the range.6–8
Clinical judgement in requesting and interpreting serum AED levels is essential. In practice, AED serum level testing has a limited role: to assess adherence to, or suspected toxicity of, a limited range of AEDs. These indications are summarized in the recently updated evidence-based epilepsy guidelines from the Scottish Intercollegiate Guidelines Network (SIGN).1 The guidelines state that monitoring may be undertaken to answer specific clinical questions; routine monitoring is not indicated. Furthermore, serum level testing is appropriate for only a limited range of AEDs: ‘Evidence supports clinically useful dose-response and dose-toxicity relationships for carbamazepine and phenytoin. These relationships do not occur with sodium valproate or any of the newer AEDs.’
|
Methods |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
Our audit standards were: ‘indication-compliant’, serum AED levels should be requested only for documented suspicion of poor AED adherence or toxicity; ‘fully compliant’, ‘indication-compliant’ requests where the relevant AED has an established dose-response and dose-toxicity relationship, e.g. phenytoin, carbamazepine, phenobarbitone.
The evidence for meeting these standards had to be either recorded in the clinical notes or readily inferred from the clinical context. The SIGN guidelines do not emphasize the necessity for trough rather than random AED levels, and hence we did not use timing of the serum AED sample as a criterion for assessing guideline compliance.
We identified all serum AED level measurement requests sent for analysis from the University Hospital of Wales to the local therapeutics and toxicology laboratory during October and November 2002. We noted the medication prescribed, the requesting team, the nature of sample (random or trough), documentation of indications and test results, its interpretation, and influence on patient management. We noted whether sufficient time had elapsed for steady serum concentrations to be reached (generally five half-lives, excluding those receiving loading doses). We also noted significant renal or hepatic co-morbidity.
Indications for requesting AED levels were grouped as follows.
Suspected toxicity
Documented suspected AED toxicity or any other clue to toxicity, e.g. diplopia, nausea.
Suspected non-adherence
Documented suspected poor AED adherence.
Post-loading
Sample the same day as the initial loading dose.
Recent fit
E.g. in emergency unit during or following seizures, but not in status epilepticus.
ITU, child/IQ
Clinical assessment difficult, e.g. toxicity in unconscious patients on ITU, adults with learning difficulties or psychiatric problems, and in children.
Not specified
No reason identified or recorded.
|
Results |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
During the two-month study period, 114 patients underwent AED level testing but relevant hospital notes were available in only 102 (68 in-patients, 34 out-patients). By coincidence, there were 114 AED level measurements in these 102 patients. Twenty-seven patients were prescribed two or more AEDs. No patient had significant renal or hepatic co-morbidity. The serum level requests were for phenytoin (n = 50), valproate (n = 27), carbamazepine (n = 22), lamotrigine (n = 8), phenobarbitone (n = 7). Additional AEDs prescribed, but for which levels were not requested, included benzodiazepines (n = 4), topiramate (n = 2), gabapentin (n = 2) and vigabatrin (n = 1). Requests were from physicians (n = 46), paediatricians (n = 30), neurologists (n = 15), neurosurgeons (n = 14), psychiatrists (n = 7), and intensivists (n = 2). Tables 1 and 2 summarize the serum AED levels results and indications, respectively.
|
|
There was written case-note documentation of test indications in 50%, and of results in 48%. Toxicity had been queried in 29 (25%), and AED adherence in 10 (9%); thus 34% of requests were indication-compliant. 16 of these requests were for patients on valproate or lamotrigine, thus only 23 requests (20%) were fully compliant. No levels were taken following a phenytoin dose change, and no requests specifically sought a phenytoin level with view to dose adjustment (such requests would have also been indication-compliant), yet 18% of patients taking phenytoin had their dose altered after the AED level result became known.
Examples of ‘non-standard’ indications in this audit included: ‘not specified’ (n = 25, 22%), ‘post-loading’ (n = 12, 10%, almost exclusively for phenytoin), and ‘recent fit’ in patients with known epilepsy (n = 19, 17%). Table 1 shows that 64 AED levels fell outside target ranges, comprising 60% of requests for those AEDs with an available target range. However, AED doses were changed following only 17 of these. In seven, the AED was reduced or stopped, all having levels above target ranges. The indications were suspected toxicity (n = 2), recent fit (n = 1), and not specified (n = 4). In eight, the AED was increased, five with levels below the target range, one within, and two above the ranges. The indications for all these AED level tests were not specified. Two other AED changes were made: a second AED added (n = 1); an AED not assayed altered in dose (n = 1).
Thus, only two of the 17 patients (12%) where AED changes were made followed guideline-compliant testing. Two results confirmed clinically suspected toxicity, but did not prompt a change in prescribed dose. Four changes were made before the results were known, or after insufficient time for steady serum concentrations had been reached. One patient had a serum level of valproate inappropriately performed for weight gain. Almost all (98%) of the phenytoin levels were random rather than trough specimens, but this was not considered a guidelines breach in this audit.
|
Discussion |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
We have demonstrated poor local compliance with national guidelines for serum AED level testing. Only 34% of requests were compliant with the SIGN guideline test indications,1 only 20% were fully compliant with the guidelines, and only 12% of those whose AED dose was changed had undergone guideline-compliant testing. We made particular effort to interpret the reasons for requesting levels, for example, inferring suspected toxicity from any documented toxic symptoms or signs, but appreciate that thoughts of busy clinicians may not always be faithfully recorded in the notes. The audit involved relatively small patient numbers, and covered hospital practice only, but nevertheless probably does represent UK practice generally, with a wide range of specialists involved in epilepsy care. The SIGN guidelines, our audit standard, had not been adopted at the time of data collection. However, they do fairly reflect current specialist opinion, and accord with previous literature highlighting the wastefulness and potential dangers of performing serum AED levels for spurious indications, e.g. ‘routinely’.5,10 Our results, therefore, demonstrate that local practice not only departs from evidence-based guidance, but also conflicts with the central message of previous authoritative publications concerning the role of AED levels in clinical practice.5,10 If reflected in practice elsewhere, this audit raises concern about widespread misuse of resources and potential clinical mismanagement from misleading or misinterpreted AED level results.4,9
Serum AED levels are most useful when applied to phenytoin, and to a lesser extent to carbamazepine or phenobarbitone. Requests for valproate and lamotrigine levels may be misinterpreted and lead to inappropriate dose changes. The SIGN guidelines suggest that valproate's short half-life makes its serum measurement almost useless, and there is no widely accepted range for lamotrigine. Yet our audit shows that this type of request is still common, comprising almost a third of requests. Even without SIGN guidance, local drug information and other literature have previously highlighted the limitations of such testing. Target ranges for serum AED concentrations are derived from trough levels; concentration changes before and after doses are small but potentially important.10,11 The SIGN guidelines do not give explicit guidance on the imperative of using trough rather than random serum AED levels, and so this was not a criterion for guideline compliance here. Had this been an audit standard, then overall guideline compliance would have approached zero.
Previous audits on the appropriateness of AED level measurement, though using different guideline standards, have given broadly similar results.10,12 It is disappointing that the practice of inappropriate requests for therapeutic drug level monitoring remains unchanged despite advice that followed these studies. Some requests probably reflect habit, since 52% of results were not even recorded in the notes. Non-standard indications included testing on the day of a loading dose (10% of requests), testing in patients with known epilepsy and recent seizure (17%), and ‘not specified’ (22%). These indications not only breach SIGN guidance but have been previously identified as typical pitfalls in AED level measurement, risking misinterpretation or prompting unnecessary or potentially harmful medication changes.5 Other indications not included in standard guidelines may be more justified: for example, children or adults with learning disabilities or psychiatric illness, and patients on intensive care units, where clinical assessment of seizure control and toxicity is difficult. One might argue, however, that SIGN covers these situations, since there was probably (undocumented) concern about AED toxicity in some, but not all.
Certain situations underline the importance of interpreting guidelines in the clinical context. For example, some clinicians may use AED level testing to document toxicity to aid communication with other clinicians managing the patient; this could explain the three cases where the AED dose was changed before the result was available. Serum AED levels are usually only helpful to check adherence if there is poor seizure control. However, frequent seizures may reflect true seizure resistance, incorrect diagnosis, inappropriate medication, progressive neurological disorder, lowered seizure threshold following alcohol exposure, or truly poor adherence. Serum levels after a single seizure may mislead, since medication may have been unavoidably missed during a fit or the post-ictal period.
General physicians requested the majority of AED levels. Neurologists and psychiatrists, despite managing the majority of difficult adult epilepsy cases locally, made relatively few requests. Patients whose AED levels were taken in breach of the guidelines risked consequent spurious management decisions. In 13 patients, the AED dose was adjusted following AED level measurements that were not ‘indication-compliant’. This led, for example, to one patient's phenytoin dose being increased from 300 to 500 mg following a random below-target range result despite optimal seizure control; admission to hospital had been for an unrelated problem. This patient became toxic, and only after considerable delay was the dose reduced. In addition, not all results from guideline-compliant requests were followed by appropriate action. For example, patients with suspected toxicity and high AED levels did not invariably have their dose reduced. In fact, clinical management was mostly unchanged following AED level measurements, even with results outside target ranges. This may indicate that clinicians did not blindly follow therapeutic ranges at the expense of clinical assessment, recognizing the well-documented shortcomings of poorly targeted AED therapeutic monitoring; alternatively, it may reflect a more random process in which some results were simply not recorded (only 48% of notes had documented results) and therefore no action followed. Whatever the explanation, these results identify an unnecessary use of resources.
In conclusion, our study highlights several shortcomings in clinicians’ indications for requesting serum AED levels locally, and in the interpretation of their results. This has implications for both patient care and resources. The shortcomings apply not only to compliance with SIGN guidelines, but also go against expert consensus views on the role of AED levels. The main points from the audit were a failure to appreciate the limited indications for requesting AED levels, the limited number of drugs with interpretable levels (carbamazepine, phenytoin and phenobarbitone), and that clinical judgement is important in analysing results to avoid unnecessary or harmful AED changes. Also, there is currently little awareness locally that sample timing (trough or random) might be important. We intend to disseminate the SIGN guidelines to colleagues caring for patients with epilepsy locally to remind them of the limitations and potential dangers of laboratory results, and will later repeat the audit.
|
Footnotes |
|---|
Address correspondence to: Dr P.E.M. Smith, The Epilepsy Unit, University Hospital of Wales, Heath Park, Cardiff CF14 4XW. e-mail: smithpe{at}cardiff.ac.uk
|
References |
|---|
|
TopSummaryIntroductionMethodsResultsDiscussionReferences |
|---|
1. Scottish Intercollegiate Guidelines Network (SIGN) Guideline no. 70. Diagnosis and management of epilepsy in adults. [http://www.sign.ac.uk/pdf/sign70.pdf] Accessed 25 April 2004.
2. Buchthal F, Svensmark O. Aspects of pharmacology of phenytoin (Dilantin) and phenobarbital relevant to their dosage in the treatment of epilepsy. Epilepsia 1960; 1:373–84.[CrossRef][Web of Science][Medline]
3. Thomson AH, Brodie MJ. Pharmacokinetic optimisation of anticonvulsant therapy. Clin Pharmacokinet 1993; 24:441–52.[Web of Science][Medline]
4. Richens A, Dunlop A. Serum phenytoin levels in the management of epilepsy. Lancet 1975; 2:247–9.[Medline]
5. Chadwick D. Overuse of monitoring blood concentrations of antiepileptic drugs. Br Med J 1987; 294:723–4.
6. Lund L. Anticonvulsant effect of diphenylhydantoin relative to plasma levels. Arch Neurol 1974; 31:289–94.
7. Turnbull DM, Howel D, Rawlins MD, Chadwick DW. Which drug for the adult epileptic patient: phenytoin or valproate? Br Med J 1985; 290:815–19.
8. Gannaway DJ, Mawer GE. Serum phenytoin concentrations and clinical response in patients with epilepsy. Br J Clin Pharmacol 1981; 12:833–9.[Web of Science][Medline]
9. Maloney TW, Rogers DE. Medical technology—a different view of the contentious debate over costs. N Engl J Med 1979; 301:1413–19.[Abstract]
10. Schoenenberger RA, Tanasijevic MJ, Jha J, Bates DW. Appropriateness of anti-epileptic drug level monitoring. JAMA 1995; 274:1622–6.
11. Richens A, Perucca E. Clinical pharmacology and medical treatment. In: Laidlaw J, Richens A, Chadwick DW, eds. A textbook of epilepsy, 4th edn. New York, Churchill Livingstone, 1992:535–40.
12. Pearce GA, Day DO. Compliance with criteria necessary for effective drug concentration monitoring. Ther Drug Monit 1990; 12:250–7.[Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||