Monograph: substitution of critical dose drugs: issues, analysis, and decision making (03-00)

Substitution of
Critical Dose Drugs:
Issues, Analysis,
and Decision Making
The National Professional Society of Pharmacists Advisory Board
Introduction
Rita Alloway, PharmDAssociate ProfessorCollege of PharmacyUniversity of TennesseeMemphis, TN Over the past few years, there has been controversy concerning the sub- stitution of generic equivalents of certain brand name drugs and the Food and Drug Administration’s (FDA’s) standards for evaluating the bioequivalence of the generic forms of these critical dose drugs. To help clarify the complex issues surrounding the use and substitution of critical dose drugs, APhA Department of PharmaceuticsSchool of Pharmacy assembled a multidisciplinary panel of experts and held a roundtable discussion on the issue. The results of that discussion and its analysis constitute this publi- cation. This publication is intended to help health care professionals and organizations understand the important issues involved in the use and substitu- tion of critical dose drugs and to enable them to make informed decisions regarding the use and substitution of these drugs in specific patients or patient The essence of the controversy involving critical dose drugs is whether the FDA’s standards regulating bioequivalence are strict enough to ensure that generic formulations of these drugs are clinically equivalent to their brand name competitors in terms of therapeutic outcomes. The brand name manufacturers of these drugs sometimes suggest that the current FDA standards are inadequate for critical dose drugs, whereas the FDA has vigorously defended those standards as being more than adequate. In addition, perhaps because they are so complex, these bioequivalence standards have been widely misinterpreted and misunderstood. Clinical Nurse SpecialistNational Institutes of HealthBethesda, MD A host of issues surround this bioequivalence controversy. The whole concept that there is a special class of drugs deemed critical dose or narrow therapeutic index is new. Not everyone agrees that such a class should even Medical Drug ProgramCalifornia Department of Health exist. There is even less agreement on the characteristic of drugs in this class, and which specific drugs should be considered a part of this class. In addition, under- standing the processes by which generic drugs are marketed and the regulations that govern how they are dispensed is important to understanding this debate.
The FDA publishes its findings regarding therapeutic equivalence in the publica- tion entitled Approved Drug Products with Therapeutic Equivalence Evaluations. This valuable resource is commonly referred to as the Orange Book. The Orange Book should be the first reference consulted in the evaluation process of the use of a critical dose drug. It is widely available in most pharmacies as well as on the This publication was developed by the American Pharmaceutical Association and supported by an educational grant from SangStat Medical Corporation. A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N Background
lication exist. APhA anticipates this list to evolve as new products are developed and brought to market.
The first step in providing a framework to aid in the This document uses the term critical dose drugs evaluation of the use of critical dose drugs in specific to refer to NTI drugs generally, but it is recognized that patients or patient populations is to define the key terms this term alone is insufficient as a foundation for the used in discussion of these issues. APhA, with the assis- development of the framework that is needed to effec- tance of its expert panel, has developed the following tively analyze the use of these products. The framework approach to define and understand the terminology is built on a foundation consisting of the definitions and used. APhA suggests that the terminology that is cur- the interactions of three specific terms: (1) critical use rently being used in the literature does not encompass drug; (2) critical dose drug; and (3) critical bioavailabil- the many complex variables that require consideration ity drug. The Venn diagram in Figure 1 illustrates how Critical dose drugs have most commonly been referred to as narrow therapeutic index (NTI) drugs.
Key Terms
The term NTI became commonly used, even though theterm used in FDA regulations is narrow therapeutic The analysis begins with the delineation of each sit- ratio. The FDA defines a drug with a narrow therapeutic uation and where it falls within the Venn diagram.
Identifying the group to which a specific drug belongs ■ There is less than a twofold difference in median must be done on a case-by-case basis. Some drug prod- lethal dose (LD ) and median effective dose ucts may fall into all three categories, and others may fall into just one or two of those categories. Others may fall ■ There is less than a twofold difference in the into different categories when used in different patients or situations. The category that a drug falls into most fre- effective concentrations in the blood, and quently will determine how it should be evaluated.
■ Safe and effective use of the drug products requires careful titration and patient monitor-ing.1 The term NTI is limiting in that it implies that the only important characteristic of the drug is the ratio of Learning Objectives
the effective dose to the toxic dose. The term wouldseem to apply, then, only to those drugs for which spe- After reading this article, health care practitioners cific drug levels may be measured. In addition, the term NTI fails to capture the importance of what effects the 1. Define and explain key terms involving the use drug may have directly on the patient.
For these reasons, the term critical dose drug has come into use.2 Critical dose drugs have been defined to 2. Examine and describe clearly the Food and Drug Administration’s standards in determin- ing bioequivalence and therapeutic equivalence ■ Requirements for blood level monitoring, for all drugs, including critical dose drugs.
■ Dosing based on body weight or other highly 3. Explain the regulation of generic drugs with ■ Serious clinical consequences of overdosing respect to both the approval process and substi- (toxicity) or underdosing (lack of effect), and ■ Steep dose-response relationship for either effi- 4. Dispel concerns and misconceptions about generic substitution of critical dose drugs.
Drugs that have typically been described as NTI or critical dose drugs include the following: lithium, 5. Provide a step-by-step framework for analyzing cyclosporine, warfarin, phenytoin, levothyroxine, carba- the use and substitution of generic equivalents mazepine, digoxin, quinidine, and theophylline. These of critical dose drugs in specific situations.
drugs have been identified in the literature as some ofthe products for which the issues described in this pub- S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G Figure 1.
Venn Diagram for Understanding Critical Drugs
Critical Use Drug
effects include the specific disease state being treated; comorbid disease states; concomitant drug therapy; the The term critical use drug was created by APhA’s patient’s age; the patient’s mental or physical condition, expert panel to aid in understanding critical dose drug or both; and the patient’s ability to adhere to the drug issues. A critical use drug is one that is used in a patient therapy. The category of critical use drugs is the broad- with one or more critical variables, which may have anadverse effect on the safety or efficacy of the drug treat- ment (i.e., a negative patient outcome). Any drug can As is evident in the Venn diagram, all critical dose become a critical use drug at any time within a given and critical bioavailability drugs fall into the critical patient (e.g., a patient with acute renal failure) or group use category. Although one could argue that all drug of patients (e.g., neonates). With a critical use drug, the therapy is critical use, the term is intended to identify emphasis is on variable patient factors as opposed to a drugs that are being used in persons who are at specific drug’s formulation or pharmacokinetic behav- increased risk (possibly transiently) for negative out- ior. For example, patient factors that may have such A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N One example is the case when imipenem-cilastatin tion). Thus, phenytoin can be considered both a critical (Primaxin®) is used in a patient with decreased renal function. Normally, the drug is used to treat difficult Most people may not consider aspirin to be a criti- infections, and it has a relatively wide therapeutic range.
cal dose drug. However, it falls into that category when However, it is renally eliminated, and with decreased it is used in large doses to treat rheumatoid arthritis.
kidney function, the drug can cause seizures. Thus, it At such doses (up to 20 325-mg tablets per day), a small becomes critical to use the drug in the correct dose and increase in dose can produce toxicity because the drug’s dosing interval to allow the body time to eliminate it kinetics at that point become nonlinear. Thus, the amount of drug consumed becomes the critical factor in Drugs used in the very young or very old, in per- sons who are critically ill, or in persons who havesignificant dysfunction of the liver, heart, lungs, brain, Critical Bioavailability Drug
or kidneys (as already mentioned) fall into the critical The third term is critical bioavailability drug. Any use category. Every health care professional shares in drug that falls into this category is also a critical dose determining whether a drug used in a specific patient, and a critical use drug. A critical bioavailability drug is at a specific time, for a specific physiologic state, con- one that is not consistently absorbed by the body. The stitutes a critical use drug. Thus, all share the variability in absorption is most often dictated by the responsibility for the analysis of drugs determined to drug’s formulation. The Sandimmune® version of cyclosporine is an example of a critical bioavailabilitydrug. Here, the liquid and the gelatin capsules go into Critical Dose Drug
solution in the small intestine poorly and the chemical A critical dose drug is one in which a small change is unsatisfactorily absorbed through the intestinal villae.
in dose or concentration results in a clinically signifi- New formulations of cyclosporine (Neoral® and SangCya®) cant change in efficacy or toxicity.2 The term critical have provided a product that is much more consistently dose drug is not sufficient for our purposes because it absorbed and utilized. These new formulations are no implies that the dose or the actual amount of drug longer considered to be critical bioavailability drugs, but absorbed is the most important characteristic to deter- cyclosporine remains a critical dose and critical use drug.
mine criticality. Patient-specific issues (as detailed A drug may be a critical bioavailability drug when above) also may play a significant role. This term administered by one route but not when administered focuses on the drug product itself—specifically formu- by another route. For example, nitroglycerin when lation and pharmacokinetic factors. In other words, a taken by mouth is destroyed by extensive first-pass critical dose drug is one for which a very precise dose metabolism and the bioavailability is quite low.
must be given in order to gain the expected therapeu- However, nitroglycerin administered sublingually or tic effect without a toxic effect. The drugs that would intravenously is not a critical bioavailability drug most often be listed in this category are those that are because these routes of administration bypass hepatic metabolism the first time through the body. Phenytoin is an example of a critical dose drug pri- marily because of its Michaelis-Menten pharmacokinetics.
Examples
The nonlinear nature of phenytoin’s metabolism leads Drugs must be categorized on a case-by-case basis to a very steep dose-response curve. Large increases in because particular ones may fall into different cate- dose may result initially in a very small change in gories depending on their use, as detailed above. For plasma concentration. However, as the patient nears the example, cyclosporine may fall into all three categories steep portion of the dose-concentration curve, a mi- regardless of the situation in which it is used. It is a crit- nuscule dose increase will lead to a disproportionate ical use drug because of the patient population in which increase in plasma concentration (and often toxicity).
it is used (posttransplant). It is a critical dose drug pri- Phenytoin is also highly protein bound both to serum marily because the variability in serum concentration is albumin and α1-acid glycoprotein, causing fluctuations frequently quite large until the function of the trans- in unbound phenytoin (i.e., that available to cause an planted organ is stabilized. Finally, as detailed above, effect) for many disease states (e.g., congestive heart when using the Sandimmune® version (primarily), it is failure, closed head injury, critical illness, malnutri- also a critical bioavailability drug.
S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G Evolution of the Issues generic) marketed after 1962, were required to prove
both the safety and efficacy of the product. Although an One of the precipitating events of the debate ANDA process existed at the time, it did not apply to concerning substitution of critical dose drugs occurred drugs marketed after 1962, so generic versions of prod- on March 26, 1997, when the FDA approved Barr ucts approved after 1962 were still required to submit Laboratories’ abbreviated new drug application (ANDA) an entire NDA, including clinical trials to prove safety for warfarin sodium over the objections of Dupont Pharma, the manufacturer of the brand name drug The 1962 legislation also provided an exemption Coumadin®. Shortly thereafter, the pharmacy literature from the NDA approval process for drugs that had been began to focus attention on the issue of NTI drugs.
marketed before 1938, as they were determined to be Then, in April 1997, the Journal of the American generally recognized as safe and effective (i.e., the Medical Association published a study claiming that generic thyroid products were bioequivalent to KnollPharmaceuticals’ Synthroid®.4 More recently (October Drug Price Competition and
31, 1998), the FDA approved a therapeutically equiva- Patent Protection Term
lent form of cyclosporine: SangCya® Oral Solution Restoration Act
(cyclosporine oral solution, USP [modified]) manufac- tured by SangStat Medical Corporation. This chemical Competition and Patent Protection Term Restoration is a generic formulation of the Novartis product Neoral® Act, commonly referred to as the Waxman-Hatch Act.5 This legislation was a congressional effort to strike a bal- These activities have led to initiatives by some ance between the two competing forces in the brand name manufacturers of these products to attempt pharmaceutical industry: the generic drug firms and to distinguish them from the generic counterparts that the innovator (pioneer or brand name) drug firms. With the FDA has concluded to be therapeutically equivalent.
the passage of this legislation, generic drug firms gained Some brand name manufacturers have characterized greater access to the market for prescription drugs, and these medications as a unique class of drugs that innovator drug firms gained certain increased periods of deserve special attention and more rigorous bioequiva- market protection through special patent extensions lence standards. The FDA has rejected this notion and and awards of periods of market exclusivity. defends its standards as being sufficient for these drugs Title I of the Waxman-Hatch Act extended the as well as any others. In addition, some brand name ANDA process to generic versions of drugs first manufacturers have lobbied for legislation at the state approved and marketed after 1962. This eliminated the level that restricts the substitution of these types of requirement that generic manufacturers duplicate drugs, an effort that has had minimal success.
expensive clinical and animal research to demonstratethe safety and efficacy of the drug product. Generic Generic Drug
manufacturers were now permitted to submit ANDAsthat relied on the FDA’s prior determinations of the Approval Process
safety and efficacy of the pioneer drug. It was recognizedthat the safety and effectiveness of the drug had beenamply demonstrated by adequate and well-controlled studies by the pioneer drug manufacturer, by the accep- The Federal Food, Drug, and Cosmetic Act (the tance of these findings by the medical community, and Act), enacted in 1938, required drug manufacturers to by the widespread use of these drug entities in patient file a New Drug Application (NDA) for each newly intro- duced drug. This NDA was required to establish the Title II of the Waxman-Hatch Act compensated the safety of the drug product. In 1962, the Kefauver-Harris pioneer companies for losses due to competition from Amendments were added to the Act. These amendments the generic companies by extending the patent terms required drug manufacturers to establish that the drug product was effective for its claimed indication(s). As a New Drug Application versus Abbreviated New result, manufacturers of any new drug (brand name or Drug Application. To submit an ANDA for a generic A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N drug, there must be a previously approved drug that is drug identified by FDA as the drug product upon which the same as the proposed drug. A generic firm must sub- an applicant relies in seeking approval of its ANDA.”8 mit information to show that the product provided in By designating a single reference listed drug as the the ANDA is the same as the previously approved brand standard to which all generic versions must be bio- name drug in terms of active ingredient, dosage form, equivalent, the FDA hopes to avoid possible significant strength, and route of administration. Additionally, all variations among generic drugs and their brand name generic manufacturers must comply with the same counterparts. Such variations could result if generic precise and very stringent FDA requirements imposed drugs were compared with different reference listed on the brand name products. They must meet the same standards for manufacturing practices, identity, There may be an instance when multiple NDAs are strength, quality, and purity. Minor differences in label- approved for a single drug product, and a product not ing are allowed to account for patent exclusivity designated as the reference listed drug and not shown to protection or manufacturer-specific modifications such be bioequivalent to the reference listed drug is actually as tablet shape, color, and so forth.
shielded from generic competition. The FDA has a Figure 2 illustrates that the only difference in process that allows a firm wanting to market a generic requirements for an NDA and an ANDA is that the ANDA version of such a drug to do so. Through this process, does not require evidence of safety and efficacy.
the product with the NDA becomes an additional refer-ence listed drug. Therefore, in some situations there Role of the Orange Book
may be more than one reference listed drug. The Orange Book is the authoritative source on the Orange Book Ratings. The therapeutic ratings con- therapeutic equivalence of FDA-approved products. It is a tained in the Orange Book are typically in a two-letter comprehensive source on all drug products with format. The first letter tells whether the product is ther- approved NDAs or ANDAs. Therefore, any drug product apeutically equivalent, and the second letter provides that is contained in the Orange Book has been fully additional information on the basis of FDA’s evaluation. reviewed and approved for safety and efficacy by the FDA.7 The ratings can be divided into three categories: A, In determining therapeutic equivalence, the FDA AB, and B. Drugs given an A rating have no known or procedure compares one generic formulation with the suspected bioequivalence problems and are considered reference or innovator formulation. This product is to be therapeutically equivalent to other pharmaceuti- referred to as the reference listed drug. A reference cally equivalent products. Those drugs given an AB listed drug is defined in FDA regulations as the “listed rating are those for which actual or potential bioequiva- Figure 2.
Requirements for a New Drug Application (NDA) versus
an Abbreviated New Drug Application (ANDA)
Brand Name
Generic Drug
S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G lence problems have been resolved with adequate bioe- for bioequivalence have been widely misinterpreted and quivalence testing. Those that are not considered misrepresented. This may be due to the fact that the therapeutically equivalent are given a B rating. requirements are quite complex and involve statistical When there are two or more reference listed drugs, the rating will be three digits. The number added on the The concept of bioequivalence is based on the rela- end indicates to which reference listed drug the generic tionship between the time course and concentration is determined to be equivalent (e.g., AB1 or AB2). achieved in blood after a dose of a drug and its expected Grandfathered Drugs and the Orange Book. Some
clinical effect. Generally, under the Federal Food, Drug, drugs that were marketed before 1938 were initially and Cosmetic Act, a drug is considered to be bioequiva- exempt from the NDA process. The FDA has taken the lent to another drug if the rate and extent of absorption position that any new dosage form of these chemicals of the generic drug are not significantly different from constitutes a new drug. Therefore, it requires the same the rate and extent of absorption of the brand name information that would be required of more recently drug when administered at the same dose under similar approved drugs. It is a common misconception that all circumstances.10 This is generally proven through the drugs marketed before 1938 are excluded from the completion of a bioequivalence study, although in some Orange Book. Examples of drugs marketed before 1938 instances more extensive testing may be required. A include quinidine, theophylline, phenytoin, digoxin, bioequivalence study is ordinarily conducted with 18 to and levothyroxine. Quinidine, theophylline, and pheny- 36 human volunteers. With a systemically absorbed toin are included in the Orange Book, which means that drug, blood levels (even from the same product) may they have gone through the FDA approval process (NDA vary in different subjects. Therefore, in a typical study, each subject receives both the innovator and the testdrug products in a randomized crossover design, and Understanding FDA Therapeutic
as a result, serves as his or her own control. Single doses Equivalence
of the test and reference drugs are administered, andblood or plasma levels of the drug are measured over Under the Waxman-Hatch Act, the FDA considers drug products to be therapeutically equivalent if they Calculating FDA Bioequivalence. The FDA then are pharmaceutical equivalents and if they can be requires bioequivalence to be calculated using three expected to have the same clinical effect and safety pro- pharmacokinetic parameters: the maximum concentra- file when administered to patients for the approved indications of use.9 Specifically, the FDA considers a
drug to be therapeutically equivalent (and therefore the area under the curve (AUC). The rate of absorption substitutable) if it meets all the following criteria: extent of absorption (the amount of drug absorbed) is a. It contains the same amount of active drug calculated by measuring the AUC. The innovator’s prod- b. It meets compendial standards for purity, be considered bioequivalent, a generic drug must similar to those of the innovator’s product (see Figure 3).
The calculations used to determine the similarity It is manufactured in compliance with Good in these pharmacokinetic parameters measuring the rate and extent of absorption involve rather complex In other words, the generic product must pass all the statistics. To be precise, the calculation requires that the same tests and have all the same qualities of the name 90% confidence interval for the ratio of the mean brand product, in addition to being bioequivalent.
generic drug fall within 80% to 125% of that of the Bioequivalence
brand name drug, using log-transformed data.
The Waxman-Hatch Act requires that a generic Practically, this confidence interval calculation shows product be proven to be bioequivalent to the pioneer that nine times out of ten, the mean response of an indi- drug before it is deemed to be therapeutically equivalent vidual (when tested in precisely the same way) would and, therefore, interchangeable. The FDA’s standards fall within the same numerical limits of the original A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N Figure 3.
Pharmacokinetic Basics: Maximum Concentration (C
Time at Which Maximum Concentration Is Reached
(T
), and Area Under the Curve (AUC) for an Innovator
and a Generic Drug
Although the AUCs (amounts absorbed) are comparable, the absorption rates (C these two products would not be considered bioequivalent under FDA standards.
test. Alternatively, there is only a 10% chance that an individual would have a mean response outside the lim- by less than 5%. In fact, the first 224 post-1962 drugs approved over the 2-year period after the Waxman- This rule often is misinterpreted or misrepresented Hatch amendments were passed, including some NTI as meaning that the mean bioavailability (rather than drugs, had an observed mean bioavailability difference the 90% confidence interval of the mean response) of between the generic and innovator products of only the generic drug must be within 80% and 125% of that of the reference product. Because the computation of a Individual Bioequivalence—Considering a New confidence interval is influenced by the study design— Approach. One criticism of the current system is that it considering the number of subjects and the intrasubject fails to take into account, or to test for, differences in variability inherent in the bioequivalence testing—the bioavailability of a drug used in certain subsets of a pop- ulation. For this and other reasons, the FDA is in the erence products must be considerably less than -20% process of considering a new approach to documenting to +25%. In fact, if the true difference is near -20% or bioequivalence. This approach is termed individual +25%, one or both confidence intervals are likely to fall bioequivalence. It allows the possibility of scaling the outside the acceptable range and fail the bioequivalence bioequivalence “goalposts” (i.e., the boundary of 80%–125%) based on variability of the reference listed S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G Figure 4.
Bioequivalence Cases
Case I: Test Product Shown to Be Bioequivalent
Case II: Test Product Not Shown to Be Bioequivalent
Case III: Test Product Not Shown to Be Bioequivalent
Case IV: Comparison of Two Bioequivalent Generic Products
Case V: Test Product Bioequivalent
A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N (innovator) drug. One possibility is that for critical dose drug may vary from its earlier versions.5,11,12 In fact, if drugs, the goalposts would always be scaled to the vari- tighter statistical intervals were used for critical dose ability of the reference listed drug. This may have the drugs, it is even possible that when an innovator firm effect of narrowing or even widening the goalposts in a reformulates its product, the product may not be bio- given instance. For example, a reference drug with a equivalent to itself.12 The FDA has stated: wide variability (e.g., Sandimmune®) would have widergoalposts (i.e., >80%–125%) than a drug with less vari- If one therapeutic equivalent drug is substi- ability (which may continue to have the 80%–125% pharmacist, and patient have FDA’s assur-ance that the physician should see the same No Therapeutic Differences Seen
clinical results and safety profile. Any dif-ferences that could exist should be no What is certainly as significant as the material greater than one would expect if one lot of described above is that no serious therapeutic differ- the innovator’s product was substituted for ences between brand name innovators and FDA- approved generic drugs have been reported.12 The FDAhas stated that no clinical data have been submitted tothe agency that would warrant narrowing the present Regulation of Generic
90% confidence interval of mean response for any drugor class of drugs.13 In fact, in a recent article citing opin- Substitution
ions of transplant pharmacists in which the majority The next step in understanding this issue focuses believed the FDA’s current bioequivalence standards on the regulation of generic substitution in the dis- were inappropriate for critical dose drugs, none of the pensing process. The process of generic drug approval, pharmacists whose transplant centers used generic as just discussed, is regulated at the federal level by drugs reported that their patients had any clinical prob- Congress and the FDA. The actual prescribing and dis- pensing of medication constitute the practice of It is reasonable to assume that if any brand name medicine and pharmacy, respectively, and are regulated manufacturer had evidence of any clinical or therapeu- at the state level by the states’ medical and pharmacy tic variability between its product and a generic one, it would report those findings to the FDA. The bottom lineis that there is no evidence that the FDA’s current bio-equivalence standards are inadequate to ensure the Product Selection Laws
clinical and therapeutic efficacy of generic critical dose In the early 1970s many states still had in effect antisubstitution laws that prohibited pharmacists from Same Bioequivalence Standards Apply to Brand dispensing a generic drug for a brand name product.
Name Products. Another important point is that brand Many of the laws had been passed 25 years earlier to name and generic manufacturers must meet the same control the distribution of substandard counterfeit requirements for bioequivalence. For example, brand drugs.15 However, in 1975, the Department of Health name manufacturers often reformulate their products and Human Services implemented the concept of max- after they conduct the clinical trials that prove the safety imum allowable costs for the reimbursement of selected and efficacy of the product. Often, the drug initially generic drugs covered by the Medicaid program.
marketed is not the same formulation as the one that Through this program, states with antisubstitution laws was used in the safety and efficacy studies.10 To gain could not participate in the Medicaid prescription drug approval for subsequent formulations, brand name program because the cost for the innovator’s brand manufacturers are required to perform the same type name product was almost always higher than the maxi- of bioequivalence tests required of generic manufac- mum allowable cost. Therefore, in the mid-1970s states turers, with the same statistical variability used when began repealing their antisubstitution laws and replac- generic drugs are compared with brand name drugs.
ing them with product selection laws that became Therefore, there is no reason to expect that a generic applicable to all prescriptions. The basis for these laws drug varies in clinical effect from a brand name drug was the Model Drug Product Selection Act that was pre- any more than subsequent formulations of that pioneer pared by the Federal Trade Commission (FTC).
S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G As a result, today, every state has some type of prod- sidered NTI drugs. Typically, the legislation prohibits uct selection law. Some states use negative formularies, pharmacists from refilling a prescription for an NTI and some use positive formularies. Some laws are drug with one made by a different manufacturer unless mandatory, and others are permissive. In mandatory they obtain documented consent from both the patient states, pharmacists must substitute a less expensive generic drug for the brand name unless the prescriber Of the three states that have passed this legislation, indicates otherwise, as required by the law. In permis- only one (North Carolina) has taken initiatives to sive states, a pharmacist may choose to substitute if the enforce it. In Virginia, the state board of pharmacy, prescriber issues the prescription in a way that permits based on an attorney general’s opinion, has determined substitution. Generally, the product selection rules that the state’s current product substitution law (which apply only when a specific product has been prescribed, consists of a check box system on the prescription form) usually through the use of a brand name. If a prescrip- is sufficient for a prescriber to authorize or prohibit sub- tion is written generically, the pharmacist need not stitution of NTI drugs. The board voted not to require refer to the product selection law. The pharmacist may any further permission from the prescriber. In Texas, dispense any product containing the prescribed drug the implementation of the statute has been held up by pursuant to a generically written prescription subject to litigation contesting the process by which the imple- Although these laws vary from state to state, they all have one common feature—that is, the prescriber Potential for Pharmacist
retains the prerogative to limit or prohibit drug product Liability
selection.10,16 The method by which the prescriber mayprohibit selection varies from state to state. For exam- In the 1970s, when states’ antisubstitution laws ple, some states require the physician to make a were being replaced with drug product selection laws, notation on the prescription such as DAW (dispense as some were concerned that pharmacists may incur written) or DNS (do not substitute). Other states may greater liability risks for engaging in drug product require the prescription to have two signature lines— selection. The FTC’s Model Drug Product Selection Act one allowing substitution, the other prohibiting it. The contained a provision designed to limit a pharmacist’s NABP Survey of Pharmacy Law contains a table sum- civil liability when substituting a generically equivalent marizing the major components of the various states’ drug in place of the brand name product. It stated: product selection laws and related examples as to howthey are similar and differ from one another.16 At least 16 states have adopted the FDA’s Orange Book as the authority (i.e., formulary) regarding what greater liability in filling the prescription by dispensing the equivalent drug productthan would be incurred in filling the pre- Specific Legislative Initiatives
scription by dispensing the brand namedrug prescribed.19 on Narrow Therapeutic Index
Drugs
Most states today have the same type of provision During the past few years, as this issue of critical contained in their product selection laws. The effect of dose drugs has manifested, some brand name manu- such a provision is that if pharmacists comply with their facturers have made legislative initiatives at the state state’s product selection laws, they may incur no greater level pertaining to NTI drugs. During 1997–1998, at liability than if they had dispensed the name brand drug least 32 states had some sort of legislation introduced or product. These provisions do not differentiate between other activity by lobbyists for brand name manufactur- critical dose drugs and other drugs.
ers to restrict access to generic NTI drugs.17 To date, With the passage of over two decades, such liabil- however, such legislation has passed in only three ity simply has not resulted. The potential that a states: Texas, Virginia, and North Carolina.18 Generally, pharmacist could be held liable for his or her role in this legislation is a supplement to the states’ product product selection is more theoretical than actual.
selection laws. Usually, the legislation defines an NTI “Nowhere in the U.S. has a pharmacist incurred addi- drug and sometimes lists specific drugs that are con- tional liability for dispensing a generic drug product in A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N accordance with a state’s generic drug substitution defendants withheld prescribed treatment. Although laws.”6 In fact, one study shows that just under 1% of all the court agreed in principle with the plaintiff (because claims against pharmacists involve generic interchange no one informed the patient about his therapeutic options), it held in favor of the defendants because the Some of the literature on the topic of critical dose patient could not show that he suffered any damages.5 drugs has suggested that a pharmacist’s potential lia- The patient was unable to prove any damages because bility in the drug product selection process is greater he could not prove there was a difference between the when a critical dose drug is substituted than when other types of drugs are substituted. As in the instance of bio- In another case, Ullman v Grant, the plaintiff had equivalence, an attempt is made to distinguish the an adverse reaction to a product that was substituted for substitution of critical dose drugs from the substitu- the drug that had been prescribed.22 The pharmacist tion of other types of drugs. There are no reported cases was found not to be liable because the court assumed of pharmacist’s liability for harm alleged to have that the prescribed product would have caused the same occurred due to the legal substitution of a generic prod- reaction. In Santiago v Barre National, Inc., although uct for a prescribed trade name product.21 This the defendant pharmacist had to settle because he failed potential liability has been overstated.
to include a warning with a dispensed generic product, It is often cited that NTI, or critical dose, drugs are the court found that it was not the substitution of a dif- involved in many claims against pharmacists. In addi- ferent product that caused the harm because the tion, a trend that shows a rise in pharmacy malpractice originally prescribed product contained the same active claims and higher verdicts and settlements is detailed.
However, this trend has nothing to do with the generic As long as generic products continue to perform substitution of critical dose drugs. It is that critical dose clinically as well as their brand name counterparts, the drugs, by their very nature, are more likely to be potential that a pharmacist’s liability may be increased involved with liability suits, regardless of whether they for substituting AB-rated generic versions of critical are substituted with generic equivalents or not. The probable reason that there have not been any malpractice cases involving the legal substitution of ageneric product, and the reason it is unlikely that there Decision Analysis
will be any of these cases in the future, is that this typeof case is very difficult for a plaintiff to prove. The plain- Algorithm
tiff would have to prove that his or her injury was “more To help health care professionals more fully under- likely than not” the result of the pharmacist’s election to stand and use the information already presented, a dispense an exact generic equivalent product rather decision algorithm has been created. This decision algo- than the brand name product ordered by the prescriber.
rithm consists of a series of statements or questions that For example, the patient would have to prove that the lead a person through the major parts of the decision- dispensed product was different in some way from the making process (see Figure 5). The following sections prescribed product. That difference then would have to explain the algorithm. In addition, after the explana- be proven to be the proximate (legal) cause of the harm tory text, three decision-making scenarios have been created to illustrate use of the algorithm.
One case that demonstrates this concept is LeJeune v Trend Services (Slip Op 96-550, June 4, 1997; 1997 La App Lexis 1541).20 The plaintiff in the case was givengeneric medications instead of the brands prescribed The medication under consideration for use is new because the insurance provider which paid for the med- and has recently been approved for marketing. The first ications determined that generics should be substituted.
step in the decision process is to compare this medica- The patient was not told by his physician, pharmacist, or tion with currently available therapies. This evaluation employer (who paid for the prescription benefit pro- consists of evaluating it for safety, efficacy, and cost.
gram) that the insurance company would pay for the Proponents of brand name critical dose drugs claim prescribed name brand if the physician indicated “dis- that the difference in price between brand and generics pense as written” on the prescription. The patient is “very, very small . . . [f]or warfarin, the difference is claimed that by being forced to take the generics, the less than $2 per prescription.”24 Although $2 per pre- S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G scription may not sound like a lot of money, for patients Under a state’s product selection law, the prescriber has receiving multiple therapies for extended periods of the option to specify the brand name or generic name times (the rest of their lives), $2 per prescription could when the prescription is written. The prescriber needs add up to thousands of dollars over a few years’ time.
to be aware of the specific regulations in the state in This is particularly important for patients whose insur- which he or she is prescribing. There are three ways for ance coverage has a maximum benefit (cap). Using the prescription to be written: (1) the prescriber speci- equally effective but less costly agents allows patients fies the brand name product with a notation indicating to be covered for the maximum length of time. From that the brand name drug is to be dispensed; (2) the pre- the perspective of the payer, lower costs for medications scriber specifies a brand name without a notation or may allow for more money to be available to fund other other indication that a brand named drug is to be dis- pensed; and (3) the prescriber indicates only the genericname of the drug.
In the first case, the generic drug would not be pre- Is the new product rated A or AB in the Orange scribable. In other words, the brand name drug would Book? If it is, then it is bioequivalent to the brand name be dispensed to the patient. Most state product selection product and can be safely and effectively substituted. If laws prohibit the substitution of a branded product if it is not, do not use this compound until it receives an the patient makes a specific request (and of course, is willing to pay for it!). In that case, the generic drugwould not be considered prescribable either.
Boxes 3 and 4
In the second case, the applicable state’s product selection law should be followed. For example, if the Is this new agent a critical dose drug? (The mater- pharmacist resides in a mandatory substitution state, he ial in the preceding text can be used to determine or she would dispense a lower-priced generic drug that whether this drug meets the criteria for a critical use, was deemed substitutable under the state’s law.
critical dose, or critical bioavailability drug.) If it is not, However, if the pharmacist is in a permissive state, he or leave this decision algorithm and evaluate the medica- she will have to consider some other factors. These fac- tion based on other merits. If it is, proceed to box 5.
tors include the patient’s needs or desires; thetherapeutic equivalence or nonequivalence of the generic drug; and safety, efficacy, and cost issues. Cost Does the patient’s current therapy consist of this issues include whether the patient’s insurance covers medication, and will that therapy be converted to the the brand name or the generic product. In addition, the alternate brand? If the answer is no, then the prescriber pharmacist should consider his or her ability to main- is initiating this medication therapy. In that case, the tain an adequate supply of whichever manufacturer’s new therapy should be initiated and monitored as clin- drug is dispensed in order to maintain the consistency ically appropriate. This situation is referred to as prescribability. If the answer to the above question is In the third case, the state’s product selection law yes, then patients will be converted (switched) to the need not be considered. However, the same factors that new therapy from existing medication. This situation is apply in the second case above (i.e., cost, safety, effi- cacy, therapeutic equivalence) would be considered.
Prescribability refers to the situation where a In today’s health care environment, cost always patient is initially being prescribed a critical dose or crit- must be considered—either in terms of the patient’s ical bioavailability drug, whereas switchability refers to insurer or the patient’s ability to pay. Generic drugs cost the situation where a patient has been stabilized on a less than brand name drugs. However, it is also impor- particular critical dose drug and is then switched to tant to remember that the cost of drug therapy include costs other than the price of the drug product itself.
The analysis for prescribability is performed by The patient will be switched to the new therapy.
examining how the prescription is written pursuant to a The analysis of the switchability of a critical dose or crit- state’s product selection law and the prescriber’s intent.
ical bioavailability drug revolves around whether one A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N formulation may be substituted for another without when critical dose therapy is switched, increased moni- concern for reduced effectiveness or increased inci- toring is automatically required; therefore, the cost dence of adverse effects. Theoretically, a switch actually benefits of switching to the less expensive drug are out- could be involved when a patient receives a different lot weighed by the increased costs in monitoring. Typically, of the same manufacturer’s drug. However, most health patients taking critical dose drugs require close moni- care practitioners would not be aware of this situation. toring regardless of whether or not one of their drug There are two parts to consider in the analysis of therapies is switched to a generic brand. switching. The first is the legality of performing the The FDA takes the position that no additional switch. The second is the issue of clinical problems clinical tests or examinations are needed when a generic resulting from the switch. Attempting to evaluate the drug product is substituted for the brand name prod- potential for problems in an individual patient is diffi- uct if they are determined to be therapeutically cult at best. The FDA bioequivalence tests state that the equivalent. This statement is based on the assumption mean response to an A- or AB-rated drug will fall within that all patients receiving these types of drugs are predictable (and clinically acceptable) limits 90% of the closely monitored anyway. “Usually these drugs [those time. If this does not happen, go to box 10.
which small changes in dose or blood level concentra- The issue of switchability is more complex clini- tion may result in clinically important changes in drug cally and requires more analysis than the situation efficacy or safety] require frequent adjustments in the involving prescribability. The first question involves dose of the drug and careful patient monitoring irre- whether such a switch is allowed under the applicable spective of whether the drug is a brand or generic drug state’s product selection law. The original prescription should be reviewed to see whether the product selectionlaw allows for the substitution of the generic drug. If it Boxes 9–11
is not allowed, the patient should be maintained on cur- Is the new agent working? If the answer is yes, con- rent therapy. If the patient desires a switch, however, tinue the new therapy (ensuring that the same the prescriber should be contacted to obtain his or her manufacturer’s product is dispensed every time). If the answer is no, evaluate the dose, patient adherence, tox-icities, and so forth. Consider switching back to the previous therapy if the patient is not tolerating the At this stage, therapy with the new critical dose drug is initiated after notifying all affected parties. If a It is best to avoid multiple switches, however, as switch is made, it is important that all parties be medication inconsistency does not allow the patient to informed by the person responsible for making the receive the maximum benefit from the prescribed ther- switch. The prescriber should be told that a switch apy. Multiple switching is the term used to describe the occurred. In addition, the patient and his or her care- situation that can happen when there are multisource taker should be advised that the appearance and name products of a certain critical dose drug. Multiple switch- of the medication may have changed. The patient also ing occurs when the patient does not consistently should be advised to notify his or her health care receive the product made by the same manufacturer.
provider if he or she notices any changes in clinical As a general rule, it is recommended that multiple effect. In the transplant environment, the transplant switching, even among AB-rated products, should not coordinator should be informed that such a change in be done in situations involving critical dose drugs. therapy has been made. (Note: Although this notifica- Although the potential for multiple switching tion would be helpful in all situations with all exists, at the present time it is not much of an issue. The medication switches, it is only truly practical when deal- reason is that most of the drugs that are considered crit- ing with critical dose drugs. For instance, a pharmacist ical dose presently do not have multiple suppliers.
would not even think about contacting a prescriber However, this issue has the potential to become more when switching a patient from one generic ampicillin to important as more manufacturers enter this market.
another, or from one furosemide to another.) The new therapy should be monitored as clinically Scenarios
indicated. Monitoring is a component of the cost of drug The following examples illustrate how to use the therapy. Some brand name manufacturers suggest that decision algorithm. The first is from the viewpoint of a S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G Figure 5.
Decision Algorithm for Substituting Critical Dose Drugs
affected parties. Monitor new therapy as dictated by clinical status of the patient.
Evaluate response at routine intervals.
patient is not tolerating the switch.
A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N prescriber, the second from a dispensing pharmacist, medicine. The patient says he does not remember.
and the third from a large insurance provider.
When asked if he wants the new medicine the patient Scenario 1: The Prescriber. M.B. is faced with a asks, “How can I be sure it’s going to work like the med- dilemma. An NTI medication that she frequently pre- icine I’m taking?! Those ‘genetic drugs’ have had some scribes is now available from another manufacturer.
She is detailed by representatives from both manufac- T.R. does not quite know where to start, but she turers and provided with marketing literature. The remembers that the pharmacy recently received a mail- literature is confusing, but the new medication is 20% ing on critical dose drug decision making. She consults less expensive on a per tablet basis. Should she prescribe that decision algorithm as a quick reference. T.R. tells the patient that the drug has been AB M.B. remembers the decision algorithm on critical rated by the FDA, which means that it is equivalent in dose drug decision making. She pulls it out and walks terms of safety and efficacy to the medicine that the through it. The first question is one of FDA equivalence patient currently takes. Also, she tells him that all rating. She calls a pharmacist friend, who tells M.B. that drugs, generic or not, are manufactured under the same stringent conditions. The patient says, “Well, I’m still By performing a literature search using Medline, not sure. See what my doctor says.” She tells the patient M.B. discovers that a small number of randomized, con- that she will ask his doctor whether he truly wanted to trolled trials have been performed with the drug in switch medications. She asks whether he can wait, or question in the same population that she treats. She whether he would like his regular medication right notes that the adverse effect profiles of the two medica- away. For the cost difference, he says he will wait.
tions are not significantly different. To date, however, The phone call to the prescriber confirmed that he there is no information on how to convert patients to wants to try this new medication and that he is com- fortable with the science behind it. He is pleased to She decides to call the company and talk directly know that it will reduce the patient’s cost as well. T.R.
with the physician overseeing the trials. The company asks the prescriber when the patient will require a sends her the material that she needs, and she decides return to the office for a checkup. He tells her not for 6 to cautiously try this new agent in some of her most sta- Scenario 3: The Large Insurance Provider. M.L., a M.B. writes a specific prescription with the new vice president of health care services, is working at his drug name (not just a generic name) and specifies the desk. One of his employees tells him that there is a new manufacturer. To ensure that the correct product is dis- critical bioavailability medication available generically pensed, she talks with the pharmacist about the newmedication and its merits and about how important it is to treat 100,000 of his covered patients. The drug is AB for the evaluation of efficacy that the new medication be rated according to the Orange Book on the Internet. Its dispensed every time. The pharmacist documents this price is significantly less than that of the name brand.
information in the patient record and highlights the How should M.L. go about making the decision as to new brand to be dispensed each time.
whether the company should pay for the generic prod- Scenario 2: The Dispensing Pharmacist. T.R. has been a pharmacist for many years. Each year it gets M.L. decides to forward the question to his formu- more and more difficult to perform the basics of dis- lary committee. He also forwards a copy of a publication pensing—particularly with so many generic versions on critical dose drug decision making to them. The of drugs available and all the insurance company for- committee reviews the literature as well as the decision algorithm. It also listens to presentations from both Today T.R. is presented with a prescription for a manufacturers. After this thorough review, the com- new medication. She has never dispensed it before, so mittee members decide that not enough studies have she consults her references. She finds that it is a com- been conducted in the target population to warrant a petitor product to the critical dose medication that the wholesale switch. However, because the product is AB patient is currently taking. She is pleased to note that rated, they approve the new product’s use as a formulary her wholesaler stocks it. She is amazed by the cost dif- item and decide to gather and analyze the usage data.
ference between the two products. T.R. asks the patient They ask both companies to return in 6 months or if the prescriber told him he was going to use a new when significant new data are published.
S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G References
The issues surrounding the generic substitution of critical dose drugs will continue as long as drug manu- 2. Barr WH. Cyclosporine: the case for expanding bio- facturers continue to exist and compete. The equivalence criteria to include measures of individualbioequivalence in relevant population subsets.
consistency of the definitions of the key terms used in Transplant Proc. 1999;31(suppl 3A):25S–30S. discussion of this issue is of the utmost importance. The 3. Sabatini S, Ferguson RM, Helderman JH, et al. Drug process by which drugs are approved and come to mar- substitution in transplantation: a National Kidney ket in the United States ensures that safe, effective, and Foundation White Paper. Am J Kidney Dis.
less expensive drug therapy is available for those patients who need it. The FDA carefully oversees the 4. Dong BJ, Hauck WW, Gambertoglio JG, et al.
complex calculations used in determining the thera- Bioequivalence of generic and brand-name levothyrox- peutic equivalence of all drugs—including those that ine products in the treatment of hypothyroidism.
JAMA. 1997;277:1205–13.
are considered to be critical dose drugs. The method 5. Pub L No. 98-417, 98 Stat 1585 (1984) (codified at 21 currently used to determine the bioequivalence of USC §355(j) [Title I]. Pub L No. 98-427, 98 Stat 1585 generic drugs is the same method required of brand (1984) (codified at 21 USC §355 (j)(4)(D)) [Title II].
name manufacturers to determine the bioequivalence 6. Molzon JA. The generic drug approval process. of subsequent formulations of their own products.
J Pharm Law. 1996;5:275–84.
There is no reported clinical or therapeutic difference 7. Mangione RA. Understanding generic drug equiva- between generic forms of critical dose drugs and their lence: generic drugs and pharmacy practice. In: US brand name counterparts with respect to efficacy or side Pharmacist’s 1999 Source to the Generic DrugIndustry. New York, NY: Jobson Publishing, L.L.C.; effect profile. In addition, to date, pharmacists have not incurred additional liability for dispensing generic counterparts of brand name products. When the sub- 9. US Food and Drug Administration, Center for Drug stitution of a critical dose drug is made, it is prudent to Evaluation and Research, Office of Information Technology, Division of Data Management Services.
The FDA’s publication titled Approved Drug Electronic Orange Book: Approved Drug Products Products with Therapeutic Equivalence Evaluations with Therapeutic Equivalence Evaluations. Available (the Orange Book) is a valuable source of information for at: http://www.fda.gov/cder/ob. Accessed July 6, 1999.
health care professionals who deal with these issues.
10. Federal Food, Drug, and Cosmetic Act, Although regulation of the substitution of generic drugs 11. Nightengale SL, Morrison JC. Generic drugs and the has been the subject of local legislative efforts, in all but prescribing physician. JAMA. 1987;258:1200–4.
one state the same rules govern the substitution of both 12. Abramowicz M, ed. Generic drugs. The Medical Letter. critical dose drugs and all other types of drug products.
Use of a decision analysis tool such as the algorithm 13. FDA position on product selection for narrow thera- provided in this monograph should lead to efficient peutic index drugs [Legal Notes]. Am J Health Syst review of any new compound. The primary decision points in the analysis process are the safety and efficacy 14. Vasquez EM, Min DI. Transplant pharmacist’s opin- of the generic product compared with the innovator ions on generic product selection of critical-dosedrugs. Am J Health Syst Pharm. 1999;56:615–21.
compound, the relative cost differential, whether the 15. APhA Special Report Evaluating Therapeutic patient is being switched or prescribed the therapy for Equivalence: A Pharmacist’s Guide. Washington, DC: the first time, how the switch is conducted, and moni- American Pharmaceutical Association; 1997.
toring for continued safety and efficacy.
16. NABP Survey of Pharmacy Law. Park Ridge, IL: With these issues in mind, all new critical dose med- National Association of Boards of Pharmacy; ications can be rapidly evaluated and patients can receive the highest level of care at the most efficient cost.
A M E R I C A N P H A R M A C E U T I C A L A S S O C I A T I O N 17. David v Goliath: a new front in the ongoing battle 21. Brushwood DB. Pharmacy Malpractice Law and between generic and brand name drug companies.
Regulations. 2nd ed. New York, NY: Aspen Law and [Citizens Against Government Waste web site.] Available at: http://www.cagw.org/publications/ 22. Ullman v Grant, 450 NYS2d 955 (Sup Ct 1982).
lookingglass/pubs.looking.David-v-Goliath.htm.
23. Santiago v Barre National, Inc., 795 F Supp 508 (D 18. Va Code §32.1-87 C; NC §90-85.27(I)(4a);Tex Rev Stat 24. Hayes AH, Baker KR, Baum P, et al. At Risk: A Multi- Disciplinary Discussion of Clinical, Economic and 19. Fink JL, Vivian JC, Reid KK, eds. Pharmacy Law Digest. 33rd rev. St. Louis, MO: Facts and Legal Issues Associated with Narrow Therapeutic Comparisons, A Wolters Kluwer Co.; 1999:CL-22.
Index Drugs. Continuing education program cospon-sored by The Physician Insurers Association of 20. Vivian JC, Slaughter RL. Drug product selection: clini- cal and legal issues for pharmacists. In: US America, Rockville, MD, and the University of Pharmacist’s 1999 Source to the Generic Drug Tennessee College of Pharmacy Knoxville, TN; 1998.
Industry; New York, NY: Jobson Publishing, L.L.C.; 25. Nightengale S. “Dear Colleague” letter. Rockville, MD: US Food and Drug Administration; January 28, 1998.
S U B S T I T U T I O N O F C R I T I C A L D O S E D R U G S : I S S U E S , A N A L Y S I S , A N D D E C I S I O N M A K I N G American Pharmaceutical Association
The National Professional Society of Pharmacists
2215 Constitution Avenue, NWWashington, DC 20037-29851 (800) 237-APhAhttp://www.aphanet.org

Source: http://www.pharmazeutische-bedenken.de/uploads/media/Substitution_of_Critical_Dose_Drugs_01.pdf