Treatment of patients with epilepsy aims for complete seizure control without intolerable drug side effects (1,2). Independent of blood drug levels, toxic effects allow titration to efficacy. However, allergic reactions, metabolically or genetically determined drug-induced illnesses, and idiosyncratic effects of drugs, while rare, may be life-threatening.

Monitoring is an attempt to detect serious systemic toxic reactions of antiepileptic drugs (AEDs) in time to intervene and protect patients. The process begins with the disclosure to patients and family members of all information required for an informed decision delivered within the framework of risks and benefits. Regularly scheduled accumulation of hematologic data, routine serum chemistry values, and results of urinalysis creates an archive (3). A rational basis for this approach was thought to reside in the Physicians’ Desk Reference (PDR) (4) and the Canadian Compendium of Pharmaceuticals and Specialties (5). Although these sources appear to define the standard of practice for many clinicians, they actually preserve observations about specific and well-defined groups of patients under close scrutiny during drug trials. Contrary to some clinical practices and these publications, evidence-based scientific criteria fail to support routine monitoring, and the resulting archival data rarely predict serious drug reactions. For example, two prospective studies (6,7) investigated the efficacy of routine blood and urine testing in patients receiving long-term AED treatment. One study (6) of 199 children evaluated liver, blood, and renal function at initiation of therapy and at 1, 3, and 6 months. Screening studies repeated every 6 months disclosed no serious clinical reactions from phenobarbital, phenytoin, carbamazepine, or valproate. Abnormal but clinically insignificant results prompted retesting in 12 children (6%), and therapy was discontinued unnecessarily in 2 children. The authors concluded that routine monitoring provided no useful information and sometimes prompted unwarranted action. A second study (7) of 662 adults treated with carbamazepine, phenytoin, phenobarbital, or primidone failed to detect significant laboratory abnormalities during 6 months of monitoring and led to the conclusion that routine screening was neither cost-effective nor valuable for asymptomatic patients. Treatment of 480 patients with either carbamazepine or valproic acid in a double-blind, controlled trial also demonstrated the lack of usefulness of routine laboratory monitoring (8).

Although habits vary in the United States and elsewhere, it is good medical practice to measure biochemical function and structural circulating elements in blood at baseline before starting treatment with a new drug (9).

Efficacy and adverse effects of drugs are the foundation for treatment decisions by physicians; however, some adverse events have led to legal actions that also have affected treatment, monitoring, and the need to document patient care. Publication of such cases occurs in several circumstances. In general, a case heard in state court will be published in the official reporters for that state only if an appellate court has produced a decision marked for publication. The same is true for some trial-level decisions made by federal courts. Publication occurs when the issues determined are deemed important or significant. For example, classic cases involving AEDs have centered on medical negligence in dosage, selection of treatment, and questions about informed consent.

The approval process for drugs used in the United States is codified in the federal Food, Drug, and Cosmetic Act of
1938, as amended in 21 USC §301 et seq (2001) and the 1962 Kefauver-Harris amendment to the Food, Drug, and Cosmetic Act; both were updated with the Food, Drug, and Cosmetic Modernization Act of 1997. The Food and Drug Administration (FDA) does not regulate drug use by physicians, who may use any licensed drug to treat patients. An attempt to restrain physicians in that respect failed (United States v Evers, 453 F Supp 1141 [ND Ala 1978]).

U.S. standards of care are derived from expert opinion, source publications, or referred articles that underlie evidence-based medicine. Other sources are textbooks and published practice guidelines, such as those from the American Academy of Neurology and the Office of Quality Assurance and Medical Review of the American Medical Association.

In medical malpractice or negligence cases, determining the standard of care for a particular treatment is of utmost importance. The standard-of-care concept extends also to the methods used to obtain informed consent and at trial is usually established by testimony from experts citing source documents or articles from referred publications. One such reference source is the PDR (10) (Table 51.1).

As with any area of law over which a state has authority, the process for determining the standard of care can differ from state to state, particularly as regards the evidentiary force of the medication package insert and information in the PDR. States tend to use these materials in one of three ways. Although the differences among these approaches are not absolute, the categorization has educational and discussion value. In the first group are states that consider the PDR and package insert as establishing the standard of care (Haught v Maceluch, 681 F2d 291, reh’ing denied, 685 F2d 1385 [5th Cir 1985]) (10). In the second group, the package insert and PDR are considered evidence of standard of care and may establish a prima facie case for negligence if a physician does not follow the prescribed directions (10). Generally, however, a physician may present evidence for using a medication outside the description in the PDR and package insert (Mulder rule and echo of Mulder; see below) (Thompson v Carter, 518 So2d 609, 613 [Miss 1987]). Mulder v Parke-Davis, 181 NW2d 882 (Minn 1970), required a physician to explain the reason for deviating from the use of a drug as specified in the PDR. Such an explanation is best included in the patient’s chart. In the third group of states, the PDR and package insert are given little credence and in some jurisdictions are inadmissible without supporting expert testimony. This is known as the echo of the Mulder rule (Spensieri v Lasky, 723 NE2d 544 [NY 1999) (10).

These discrepancies in the handling of medical malpractice issues illustrate why it is critical to know local, regional, and national standards of practice and the idiosyncrasies of applicable law in a jurisdiction, as well as why a physician must diligently document the rationale for action in a patient’s medical record.

Issues of informed consent have also required adjudication. In Serigne v Ivker (808 So2d 783 [La App 4th Cir 2002]), the plaintiff alleged that informed consent had not
been obtained because teratogenicity had not been disclosed. The court found (a) that the plaintiff had failed to establish a connection between malformations and phenytoin and (b) that informed consent did exist. In Spano v Bertocci, a plaintiff claimed lack of informed consent based on nondisclosure of the teratogenic effects of valproic acid. At the trial and appellate court levels, informed consent was deemed to have been obtained because of the plaintiff’s previous knowledge of the danger of valproate use during pregnancy.

The landmark decision of Harbeson v Parke-Davis (656 P.2d 483 [1983]) illustrates the diligence required in providing information for patients with childbearing potential. A woman who delivered children with fetal hydantoin syndrome claimed failure of informed consent causing wrongful birth and wrongful life. The court stated that a physician had a duty to “exercise reasonable care in disclosing ‘grave risks’ of (any) treatment” advocated. The physician had failed to search the literature, which would have uncovered the dangers of using phenytoin during pregnancy and would have allowed the physician to inform the patient of the risks.

Serious skin reactions, including Stevens-Johnson syndrome, have also raised issues of informed consent. Shinn v St. James Mercy Hospital (675 F Supp 94 [WDNY 1987]) centered on the claim that serious skin reactions to phenytoin had not been disclosed. The court decided that all adverse effects need not be disclosed to a patient, only the most common. In addition, given the patient’s medical circumstance, treatment would not reasonably have been declined even if adverse effects had been delineated. Similarly, another court found that warnings in the PDR and package insert diminished the “danger-in-fact” of the medication: “… no reasonable trier of fact could conclude that this … medicine is unreasonably dangerous per se” (Williams v Ciba-Geigy, 686 F Supp 573 [WD La 1988]).

Documentation can be critical. A patient treated with phenytoin suffered hepatotoxic reactions, and the court originally found for the plaintiff. That decision was overturned on appeal, the appellate court stating, “Viewing the record … the skill and care exhibited by defendant’s physicians’ diagnosis and treatment were marked by devoted diligence and attention and were wholly consistent with the professional skill … employed by other physicians in treating and controlling … the complex disease of epilepsy” (Fritz v Parke-Davis, 152 NW2d 129 [1967]).

Errors in the use of AEDs that amount to negligence have resulted in legal action. When a patient who was to take Dilantin 500 mg per day received a prescription for 500 mg three times a day, judgment was for the plaintiff (Hendricks v Charity Hospital of New Orleans [La App 1987]). In Martin v Life Care Centers of America (No. 95-4124-B, 117th Judicial Dist Ct, Nueces County, TX, April 1998]), high plasma levels of Dilantin were associated with a patient’s death, resulting in judgment for the plaintiff. One court found for the plaintiff in a case of failure to diagnose pancreatitis from the use of valproate (Pester v Graduate Hospital, No. 87-05-00357, Court of Common Pleas, Philadelphia, PA, Oct 1992).

Serious idiosyncratic drug reactions do not depend on dose and by their nature are unpredictable (11, 12, 13). All organs are affected, the skin most commonly (Table 51.2). Established AEDs, used in millions of patients, are known to cause agranulocytosis, aplastic anemia, blistering skin rash, hepatic necrosis, allergic dermatitis, serum sickness, and pancreatitis. Newly available drugs, used in many fewer patients, have caused allergic dermatitis and serious skin reactions (Table 51.2). With the exception of reactions to felbamate, other serious reactions have yet to be reported with any alarming frequency.