Identifying "Triggers"

Several tools and tips are available to help healthcare professionals establish that an ADE has occurred. "Triggers" are surrogate markers suggesting that some unfavorable outcome has occurred with a drug therapy. One of the simplest ways to identify triggers is to review a patient's medication profile and search for any drugs that were recently or abruptly discontinued. Although there are many reasons why a drug might have been discontinued, at least one possibility is that the patient experienced an ADR.

Clinicians also can review a patient's medication profile and look for any new allergies that have been added to the patient's history. For example, if someone is listed as "NKDA" (no known drug allergies) upon hospital admission, but then a new drug allergy is added to the patient's chart during the hospitalization, that would suggest that an adverse event may have occurred, such as an allergic reaction to a prescribed drug.

Similarly, a review of the patient's medication profile may reveal new orders for medications that are antidotes or alternative therapies, which also may suggest that an ADR has occurred. Consider, for example, a patient who is listed as NKDA and who receives the drug piperacillin during a hospital stay. A review of the patient's record reveals that at some point, the piperacillin was discontinued and the drug aztreonam was initiated. The "trigger" here is the switch from one antibiotic to another. Aztreonam is a beta-lactam antibiotic of the monobactam class, and it is used almost exclusively in patients with a beta-lactam antibiotic allergy. In aztreonam, the primary allergenic determinant -- the intact beta-lactam ring -- is "split" by a carbohydrate chain, disallowing for cross-reactivity in patients who are allergenic to that particular intact ring structure.

Table 3 lists some known drug "triggers" that, when added to a patient's orders during a hospitalization, suggest that an ADR has occurred.

Table 3. Drug "Triggers"

Healthcare workers often are hesitant to report potential ADRs if they are not absolutely certain that the drug was the cause of the adverse outcome. Although probability is sometimes difficult to determine, it is important to realize that a 100% certainty is more the exception than the rule. Thus, a degree of uncertainty should not deter the reporting of any event with some type of temporal relationship to a specific drug therapy.

Still, clinical judgment alone may not be sufficient for a clinician to decide whether a drug was the likely cause of a certain adverse event, and in fact, reliance on clinical judgment alone has shown poor correlation with the correct identification of ADRs. Fortunately, there are tools available to augment the clinician's judgment call.

One of the most widely used and validated tools to assess the likelihood that a specific drug is the cause of an ADR is the Naranjo algorithm^[23] (Table 4). This algorithm uses a series of 10 questions, with numerical values assigned to each answer, to determine an overall score for drug-related probability.

Table 4. Naranjo Algorithm*

*Adapted from Naranjo CA, Busto U, Sellars EM, et al. Clin Pharmacol Ther. 1981;30:239-245.

The total score is calculated by adding the scores for each of the 10 questions, and these final scores indicate the likelihood that the event in question was an ADE:

Definite ≥ 9
Probable 5-8
Possible 1-4
Doubtful <>

Along with a probability assessment, an ADE report should provide a meaningful description of the event and its outcome, as well as any contributing or competing factors, and, of course, the patient's medication history. Contributing and competing factors may be present when concurrent disease states or conditions make a patient more susceptible to an ADR. The next case illustrates this concept:

Mr. B is a 68-year-old man with a history of colon cancer. He also has a past medical history that is significant for heart failure and atrial fibrillation. His regular medications include dofetilide 250 mg every 12 hours for atrial fibrillation, furosemide 40 mg once daily, and warfarin 5 mg daily. He also recently received treatment for his colon cancer with 5-fluorouracil (5-FU). A week after this last treatment, he developed significant diarrhea. He continued to take all of his other medications as prescribed. He presents to the emergency department after experiencing new palpitations, and he also exhibits unusual bruises on his extremities. On admission, he is noted to have an INR (international normalized ratio) of 7.5 and a potassium level of 2.5 mEq/L. An EKG reveals a QTc interval of 686 msec with ventricular polymorphism.

After reviewing and analyzing this case, we can conclude that all of Mr. B's conditions are drug-induced. What are the factors involved, and how do they exacerbate each other?

1. Diarrhea is an adverse effect of 5-FU.
2. Diarrhea can lead to hypokalemia and other electrolyte imbalances.
3. Furosemide can lead to hypokalemia, and in the setting of concomitant diarrhea, this effect is augmented.
4. Hypokalemia is a predisposing factor to increase QTc prolongation in the setting of a patient's use of antiarrhythmic agents such as dofetilide.
5. 5-FU also has an interaction with warfarin, increasing INR levels.

Taking the First Step: Reporting ADEs

As discussed earlier, ADEs are a significant patient safety and public health problem. However, it is possible to use these incidents to improve patient care; the key is reporting them. Reporting is critical not only for postmarket surveillance of drugs, but also for developing a pharmacoepidemiologic surveillance program, at both the institutional and national levels.

It is important to keep in mind that although clinical trials have highly regulated methodologies, they also have significant limitations in defining a drug's safety profile. Because of the sample size used in clinical trials, the only adverse reactions that are usually discovered are those that occur frequently.

For example, if a study enrolls 10,000 patients and not a single patient has a particular serious reaction, one can still only be 95% confident that the likelihood of that reaction occurring in that sample size is less than 1 in 3333.^[16] In order to raise the threshold of safety to a confidence level of 99%, the study would need to involve 46,000 patients^16]; however, clinical trials generally do not enroll that many patients, and even some of the largest clinical trials enroll only 5000-10,000 patients.

An example of how serious drug reactions were not detected in clinical trials can be seen with the drug bromfenac. Bromfenac is an NSAID that was marketed for acute pain and inflammation. During the postmarketing experience, however, a number of adverse events were reported involving the use of this drug, including reports of fatal liver toxicity. Unfortunately, serious liver damage occurred in about 1 in 20,000 patients who used bromfenac for more than 10 days. However, the incidence of death due to hepatotoxicity was below the practical detection limit of most clinical trials. In order to have reliably detected serious liver toxicity at the clinical trial level, the researchers would have had to expose 10,000 patients to the drug.^[17] After weighing the postmarketing reports of hepatotoxicity, the FDA decided only to issue a recommendation that bromfenac be used for less than 10 days. This labeling change, however, did not stop prescribers or patients from using this drug for longer periods of time. Eventually, the FDA concluded that the risk for hepatotoxicity outweighed any benefit of bromfenac, given the availability of other drugs in the class that were not associated with such serious adverse effects. As a result, the FDA and the manufacturer (Wyeth-Ayerst) agreed to withdraw the product from the market -- after it had been used by 2.5 million patients.

The effectiveness of postmarketing surveillance and ADE reporting can also be seen in the cases of several other drugs that were removed from the market after unexpected adverse reactions were reported (see Table 2). None of these reactions had been evident during clinical trials of these drugs.^[17]

Table 2. Scope of Patient Exposure to Drugs Removed From Market

For healthcare professionals, it can be a challenge to know what, when, and where to report an ADE. The FDA's MedWatch program^[18] was created to collect reports of serious adverse events. The agency has defined an ADE as "serious" when one or more of the following patient outcomes occur as a result of the ADE:

• Death;
• Life-threatening condition;
• New hospitalization;
• Prolonged hospitalization;
• Permanent disability or damage; and
• Congenital anomaly or birth defect.

The FDA's definition also includes ADEs that required some type of intervention to prevent permanent impairment or damage to the patient.

Most individuals prefer to report an ADE to the FDA's MedWatch program, which provides a convenient online form with directions on how to file the report.^[19] Some clinicians may feel more comfortable reporting the event directly to the drug manufacturer by contacting the company's adverse event department, usually by phone. Regardless of how these "serious" events are reported, the FDA and the pharmaceutical companies are required to share these reports with each other. This allows them to determine whether further action needs to be taken, such as updating the drug safety information for the drug involved. The key is for healthcare providers to use whatever mechanism is most comfortable for them to report the adverse events that they see in their practices.

In addition to reporting ADRs to the FDA and drug manufacturers, internal reporting systems within medical institutions can also have a significant impact on improving patient care. Although filing institutional reports is usually voluntary, the Joint Commission on Accreditation of Healthcare Organizations (JCAHO) does require that hospitals have procedures and methods in place so that employees can report ADEs. In fact, the establishment of such a reporting system is a key element in hospital accreditation evaluations. The JCAHO medication management standard MM.6.20 also requires that the hospital respond appropriately to actual or potential ADEs that are reported.^[20]

The key to producing effective ADE reports is knowing which incidents to report and what information to include. Many times, healthcare workers believe that ADRs are limited to events such as rashes and hives caused by antibiotics, or other characteristic events that involve the immune system. As mentioned earlier, these "type B" reactions are unpredictable because they are specific to individual patients, and the mechanism of what causes these reactions is not clearly elucidated. Therefore, due to their idiosyncratic nature, there is little that can be learned from these events that will improve patient outcomes, other than identifying a given patient as allergic to that specific drug, to prevent future occurrences.

In fact, type B reactions comprise only about 10% to 20% of all ADRs, so why do they tend to be reported more than type A events? One reason is that many healthcare professionals view type A events -- those related to a drug's pharmacologic effects -- as merely "side effects." However, these types of ADRs make up the most frequently occurring adverse reactions, and they are, by definition, generally predictable and dose dependent.^[21] Due to these factors, they represent the most preventable ADRs.

The following scenario illustrates how certain "side effects" can be classified as adverse reactions:

Mrs. M. is a 70-year-old woman admitted to the hospital with dehydration and anorexia. She has a history of diabetes mellitus and osteoporosis. Her medications prior to admission included insulin glargine 50 units at bedtime, glyburide 2.5 mg twice per day, and alendronate 70 mg once per week. She is believed to have a gastrointestinal viral syndrome and has a poor oral intake of food, yet her oral insulin glargine and glyburide are still being administered. She is found at 6 am to be diaphoretic with an acute change in mental status, and her glucose level is found to be 10 mg/dL. She is treated with dextrose 50% and administered an infusion of dextrose 10%. Her diet is advanced because she will require a continuous intake of carbohydrates to prevent repeat events of hypoglycemia throughout the day. Meanwhile, insulin glargine and glyburide are temporarily discontinued.

Should this case be reported as an ADE? Is the patient's hypoglycemia "just a side effect" or potentially more serious? According to the WHO definition, this is indeed an ADR. Criteria established by the American Society of Health-System Pharmacists (ASHP) also confirm that this patient has experienced an adverse event. These criteria^[22] state that an ADR is a reaction that is potentially life-threatening, causes permanent damage or even death, or requires intensive medical intervention (similar to the FDA MedWatch definition of a serious ADR). The ASHP definition also includes reactions to drugs that require a change in drug therapy (usually discontinuation or a dose reduction) and/or initiation of a specific treatment to prevent further harm to the patient.

What Constitutes an Adverse Drug Event?

"Adverse drug event" is a broad term that encompasses a variety of incidents with problematic outcomes that occur within the context of drug therapy. Simply put, an ADE is any unfavorable event associated with the use of a medication. One of the most commonly accepted definitions is based on the International Conference on Harmonization Guidelines^[13]: "any untoward medical occurrence in a patient or clinical investigation subject administered a pharmaceutical product and which does not necessarily have to have a causal relationship with this treatment." The definition continues: "any unfavorable or unintended sign, symptom, or disease temporally associated with the use of any dose of a medicinal product, whether or not considered related to the medicinal product."

Although the definition of an ADE may seem broad, it is designed to be very inclusive. Physicians, pharmacists, nurses, and consumers often mistakenly believe that an ADE must have a definite temporal relationship with a drug therapy, or that a drug must be the sole cause of the event; this is not true, and this misunderstanding may partially explain why many ADEs are not recognized or reported. Causality and probability that an event is related to drug therapy will be discussed further.

There are 5 different categories of ADEs:

1. Adverse drug reactions
2. Medication errors
3. Therapeutic failures
4. Adverse drug withdrawal events
5. Overdoses

Adverse Drug Reactions

Causality comes into play within the first category, adverse drug reactions (ADRs). In these cases, there is a defined temporal relationship between the adverse event and the administration of the drug. A commonly cited definition of an ADR is delineated by the World Health Organization (WHO) as "any response to a drug that is noxious and unintended, and that occurs at doses normally used in man for the prophylaxis, diagnosis, or therapy of disease."^[10]

ADRs generally are assigned to 1 of 2 subgroups. Events that fall within the first subgroup are known as type A reactions: an "A"ugmentation of a drug's primary or secondary pharmacologic effect. A simple example of a type A reaction is coagulopathy with warfarin. The primary therapeutic effect is to interfere with the hepatic synthesis of vitamin K-dependent coagulation factors, but when there is too much inhibition, unwanted hypoprothrombinemia and bleeding episodes may result.

Drugs also may exhibit adverse reactions due to a secondary pharmacologic effect. For example, this is sometimes seen with fluoroquinolones. Their primary pharmacologic effect is to inhibit certain bacterial enzymes, exerting their antimicrobial activity. However, fluoroquinolones can also inhibit gamma-amino butyric acid (GABA), an inhibitory central nervous system neurotransmitter; inhibition of this neurotransmitter is believed to be the mechanism of fluoroquinolone-induced seizures, which represent a rare adverse reaction to this class of drugs.

The other general subgroup of ADRs encompasses type B reactions, which are considered "B"izarre, or idiosyncratic, in nature. Such reactions are not related to a drug's primary or even secondary pharmacologic effect. The best examples in this group are allergic reactions such as rashes, hives, and systemic anaphylaxis.

Table 1 offers a comparison of the 2 types of ADRs.

Table 1. Descriptions of Adverse Drug Reactions (ADRs)

Medication Errors

Medication errors represent another classification of ADEs. Medication errors are defined as "unintended acts (either of commission or omission) resulting in actual or potential harm to a patient, or an act that does not achieve its intended outcome when dealing with drug therapy."^[14] Medication errors of commission are generally blatant even though usually unintended; for example, a patient who is supposed to receive the drug Lamictal (lamotrigine) for seizure control instead receives the drug Lamisil (terbinafine) for onychomycosis. This type of error often occurs when different drug names look or sound alike.

Another type of medication error of commission is that which results from unclear or misinterpreted handwriting. An example is the patient who is supposed to receive 1 mg of a drug but instead receives 10 mg when the order for 1.0 mg is misinterpreted as 10 mg, because the decimal point was either overlooked or poorly written. Whether such an error results in patient harm may depend on when it is discovered during the medication-use process system. (When such errors are corrected before any patient harm occurs, they are often described as "near misses.")

Medication errors of omission are those that occur when a clinician fails to detect an important factor in a patient's drug therapy regimen, such as a clinically significant drug interaction, allergy history, or disease state contraindication. Errors of omission also may occur when important drug information is not adequately conveyed to prescribers or patients. All of these omitting factors have the potential to result in patient harm.

Consider, for example, the case of a pharmacist who dispenses the fertility drug clomiphene citrate to a young woman; within the next 3 months, the pharmacist dispenses to this same patient the drug misoprostol to prevent gastrointestinal distress caused by a nonsteroidal anti-inflammatory drug (NSAID) that she is taking. The error here is that misoprostol is labeled by the US Food and Drug Administration (FDA) as a category X pregnancy risk drug, which is contraindicated for women who are pregnant and or who are planning to become pregnant. A woman who is taking clomiphene citrate is clearly trying to get pregnant, and the misoprostol could cause a spontaneous abortion or could result in congenital defects in the baby, representing a most egregious error. In this scenario, the errors of omission are the pharmacist's failure to conduct a full prospective drug-utilization review and the failure to communicate warnings associated with the drug's labeling to the patient.

Therapeutic Failures and Adverse Drug Withdrawal Events

The third and fourth types of ADEs occur when drug therapy is "not given" or "not optimally given." In a therapeutic failure, suboptimal amounts of a medication are given to a patient, and the patient's condition fails to stabilize or symptoms become even worse as a result. For example, certain antibiotics bind to the minerals found in some other drugs, and if taken together, the amount of antibiotic available for absorption will be lowered. This may result in a treatment failure.

Sometimes a patient may not fully understand the importance of the drug regimen and may not take the drug as prescribed or may discontinue therapy altogether; this often occurs in patients with chronic disease states. Medication nonadherence is a major factor in therapeutic failures.

In contrast, an adverse drug withdrawal event (ADWE) occurs when a patient on long-term drug therapy experiences a withdrawal syndrome after discontinuing taking the drug, especially when this is an abrupt change. Examples of ADWEs include rebound hypertension in patients on long-term clonidine use, as well as agitation and/or seizures in patients abruptly discontinuing long-term benzodiazepine therapy.

Drug Overdoses

The final class of ADEs includes drug overdoses, whether intentional or accidental. These events differ from adverse reactions in that the doses are not "those that are normally used to treat disease." Accidental overdoses occur most commonly in 2 patient populations at opposite ends of the age spectrum: the very young and the very old. In these instances, the root causes generally are medication compliance problems and patient or caregiver knowledge deficits about the drugs involved.

Introduction

Adverse drug events (ADEs) are a significant cause of morbidity and mortality. In the United States, the annual cost of these events has been estimated at nearly $150 billion.^[1] Within the past decade, ADEs have been the focus of many studies in patient safety and quality control, and have been recognized as a top priority in safety efforts due to their iatrogenic nature.^[2-4]

Epidemiologic studies regarding the scope of these adverse events have found:

• About 3% to 28% of all hospital admissions are related to ADEs^[5-9];
• An estimated 5% to 20% of patients experience an ADE during their hospitalizations^[10];
• Elderly patients (> 65 years) are 2.5 times as likely to have an ADE that requires an emergency room visit compared with the general population, and they are 8 times as likely to require hospitalization^[11];
• Drugs with narrow therapeutic indices and/or those that require outpatient therapy monitoring account for 41.5% of all drug-induced hospitalizations^[11]; and
• Nearly two thirds of adverse drug reactions that required hospitalizations were considered to be potentially preventable.^[12]

The potential for preventing ADEs is a key element of efforts to improve patient care.