Generic Drugs: Are they Equivalent?

With healthcare costs continuing to rise, generic drugs are looking more attractive than ever. The prospect of getting the same drug at a lower cost is tempting to anyone with a large drug bill — patient or insurer alike. The savings are massive: Lipitor lost patent protection last month — it was a $10 billion drug, and the generic versions are priced at a fraction of the original cost. In 2012, Plavix and Seroquel, two other blockbusters, will lose patent protection too — that’s another $10 billion in drug costs that will shrink. This “patent cliff” will shrivel about $255 billion in worldwide patented drug sales over the next five years. If you’re taking a prescription drug and not already on a generic, you probably will be soon. And depending on where you live, you may be automatically switched to a generic version of your prescription drug as soon as it’s available.

Pharmacists are responsible for most of the switches from brand to generic drugs. In Ontario, where I work, regulations specify which drugs and brands may be automatically substituted — that is, without patient or prescriber consent.  This doesn’t mean a  lack of transparency, however, so I spend a lot of time speaking with patients about generic drugs. Misconceptions are common, ranging from manufacturing standards (“they’re weaker!”) to efficacy (“the drugs don’t work!”). I’ve seen a number of questions and comments about generic drugs in the comments section here at SBM as well. So today’s post is an overview of the science of evaluating generic drugs. Specifically, I want to review the concept of bioequivalence, the confirmation of which assures us of the interchangeability of different drugs — that is, one can be substituted for another.

What is a generic drug?

What is referred to as a “generic” drug may vary by country, and be influenced by both medical practice and by regulatory requirements. The most common definition is that used by the World Health Organization:

A generic drug is a pharmaceutical product, usually intended to be interchangeable with an innovator product, that is manufactured without a licence from the innovator company and marketed after the expiry date of the patent or other exclusive rights.

Generic products may also be called “multisource” products. And you’ll often see a drug’s chemical name referred to as the “generic name” or the “non-proprietary” name,  which I’ve described can lead to confusion among consumers who may only know their prescription by the brand name alone.

The Active Pharmaceutical Ingredient

To understand the scientific basis of generic drug evaluations, it’s necessary to understand some key concepts. The first one is the active pharmaceutical ingredient or API. In Lipitor, for example, the API is atorvastatin, or to use its full chemical name, (3R,5R)-7-[2-(4-fluorophenyl)-3-phenyl-4-(phenylcarbamoyl)-5-propan-2-ylpyrrol-1-yl]-3,5-dihydroxyheptanoate. The API is the chemical that has the desired biological effect.  There may be a dozen ingredients in a tablet, for example, but the API is the ingredient we’re interested in. It’s the API that will allow us to generalize data and studies with the drug, linking the original bench science and preclinical research, to the tablet dispensed by the pharmacy — it’s the same chemical.  The fact that drugs have an API allows generic drugs to be marketed, because when we compare generics, the API is the same. In contrast, consider the scenario of an herbal remedy. A single tablet of 100mg of a raw herb might contain hundreds of different chemicals. If there is no known API or standardized active ingredient, we cannot compare between brands, or assume that clinical trials with one brand are relevant to any other product, because we have no idea which ingredient is actually having an effect, and if any other version has that same ingredient (or combination of ingredients.)


Most dosage forms (e.g., tablets and capsules) are designed to deliver the API to the site of action. Unless it’s a drug that acts directly on the lining of the gastrointestinal tract, we rely on the circulatory system to carry the drug to the site of action in the body.  Bioavailability refers to the amount of drug that, once ingested, reaches the bloodstream. Bioavailability is evaluated based on two measures — the rate of absorption, and the extent of absorption:

Blood plasma level curve for a typical single dose of an oral drug


Look at this graph carefully. Let’s assume a dose of a drug is taken by mouth at time=0. A dosage form, such as a tablet, needs to disintegrate, and then the API needs to dissolve. It then must be absorbed through the gastrointestinal tract, where it will move into in the blood. Sequential blood samples are taken (say, every 30 minutes) and measured for the concentration of the drug. The point measurements are then connected, resulting in the smooth line you see. At first, the drug is absorbed rapidly, peaking at concentration Cmax at time tmax. The drug then starts to be eliminated from the blood — it could be metabolized, perhaps by the liver, and/or be filtered out of the circulation, by the kidneys. We’re then able to calculate the rate of absorption, described by Cmax/tmax, and the total extent of absorption, which is the area under the curve (AUC), calculated using integral calculus.

Studying the physiological actions of the API, and relating it to the concentration time profile allows us to understand if there is a maximum tolerated concentration (MTC) and if there’s a certain blood concentration where the drug’s action seems to start and wear off, termed the minimum effective concentration (MEC). These values are estimates, based on the understanding that there is a relationship between the drug’s concentration in the bloodstream, and the amount of that same drug at the site of action in the body. These measures will guide how the drug is eventually manufactured: We want each dose to predictably and reliably follow the same curve. That means standard production methods, and the use of fillers, diluents and other excipients, all of which serve to ensure that there is little to no variation tablet-to-tablet or dose-to-dose.  This is done with the intention of also minimizing interpatient variability, which is the variation seen between different patients given the same dosage form.

There may be other ways that the formulation can be varied. When the MEC is very close to the MTC, greater precision in per-tablet accuracy may be necessary. Further, we can vary shape of the curve, by doing things like enteric-coating the tablet, or making a sustained-release version that releases the API more slowly. The final absorption curve of any pharmaceutical preparation will be a product of both the physical characteristics of the drug (i.e., the intrinsic chemical properties) and the dosage form (i.e., the pharmaceutical properties).

Generic manufacturers that want to duplicate a drug coming off patent then have a few challenges. First, manufacture the complicate chemical structure that is the API (or find someone else who can make it for you). Second, package it in a dosage form that resembles the patented drug.  Third, show that the new generic shares the same absorption curve as the branded version. That is, they must show that two product are bioequivalent.


We now want to compare a branded drug, one that’s been on the market for years, with a new generic. There is a single fundamental assumption that underlies the comparison:

Two products are considered equivalent when the rate and extent of absorption of the generic drug does not show a significant different from the rate and extent of absorption of the brand drug, when administered at the same dose under similar experimental conditions.

So let’s unpack this. We don’t need to know if the tablets look alike, or have different fillers. We also don’t need to do clinical trials with the generic drug. If we can demonstrate that the API is absorbed at the same rate and extent as the brand drug, then we can declare two products to be bioequivalent. That is, taking one or the other results in the same “body exposure” to the same drug. Assume we want to compare Drug A and Drug B. We will take a group of normal, healthy adults and put them in standardized conditions — usually fasting overnight and then giving a drug on a empty stomach or with a standardized meal. We’ll give Drug A to a patient, take serial blood samples, and then calculate Cmax/tmax and AUC. Later, after the drug has been fully eliminated from the body, we’ll repeat the process in the same patient, under the same conditions, with drug B. This may be done in 20+ patients. We’ll get a graph that looks like this:

Typical concentration-time graph for two drug products


Are the two curves super-imposable? Not quite, in this example. There are slight differences in the extent of absorption. Are these two drugs bioequivalent? Here’s where the statistics come in. We apply statistics to evaluate if the curves are meaningfully different. To do this, we need to make some assumptions about what we will accept as a meaningful difference. Does a slight difference in the rate or extent of absorption make a clinical difference?  Most regulators worldwide have decided that a 20% variation is generally not clinically significant.

Two versions of a drug are generally said to be bioequivalent if the 90% confidence intervals for the ratios of the geometric means (brand vs. generic) of the AUC and Cmax fall within 80% and 125%. The tmax (brand vs. generic) must also be comparable — and there should not be any significant differences between different patients.

For practical purposes, generic versions of branded drugs have AUC and Cmax ratios that are very close to one.  With significant variation in either value, it would be unlikely for the confidence intervals to lie withing the 80% to 125% range. For the sake of keeping this post short I’ll leave a more detailed discussion of the statistics to theinterested reader.


Myths, Misconceptions, and Controversies

Generic drugs are manufactured differently from branded drugs, and branded drug manufacturers use better processes

If the dosage form releases the same drug with the same concentration/time profile, then any minor manufacturing differences are irrelevant to the pharmacological activity. Regulators have established standardized manufacturing practices and standards, termed good manufacturing practices, that all manufacturers, brand or generic, must adhere to. Final products must meet the same product quality standards as well.

Biosimilars are bioequivalent

Biosimilars refers to follow-on versions of antibodies, proteins and other patented biological drugs, like growth hormones. Biological drugs are very large molecules that are structurally more complex that the relatively simple, smaller molecules in most tablets and injections. Their complexity means that a simple bioequivalence evaluations may be insufficient to fully describe a drug’s action. Manufacturing processes may involve recombinant DNA technology, and can take place inside living cells, leading to a final dosage form which cannot be fully characterized as a single defined API. As some of the biologic drugs approved over the past few decades have lost patent protection, the challenge of verifying both pharmaceutical but also clinical equivalence of “generic” biologicals has become an evaluation challenge for regulators. So if they’re not completely bioequivalent, do they have the same effects in patients? That is, are they clinically and therapeutically equivalent? While approaches between countries may vary, many are being cautious and requiring more detailed comparisons, including clinical safety and efficacy trials to adequately demonstrate that “generic” versions of biologic drugs have the same end effects. As regulators gain more experience evaluating these products, expect the testing standards to evolve.

Bioequivalence studies are always necessary to establish interchangeability

For some dosage forms, bioequivalence studies may be inappropriate or impractical. Consider ophthalmic eyedrops. These products are administered at the site of action, and don’t rely on systemic absorption. It may be more appropriate to verify equivalence by comparing physicochemical properties including pH, viscosity, and osmolarity/osmolality. Where all relevant parameters are evaluated to be comparable, regulators will deem generic versions interchangeable based on these parameter alone.

Generics need to be tested in sick patients, not healthy volunteers

The intent of a bioequivalence comparison is to determine the differences in formulation between two drug products. Healthy standardized volunteers are used in bioequivalence studies to control for patient differences, meaning that any differences in bioavailability will be the result of drug formulation issues, not patient issues. In order to isolate drug formulation effects, it is necessary to hold everything else as constant as possible. If formulations are consistent, the rate and extent of drug absorption will not differ, and consequently there should be no differences in the pharmacologic effects of the drugs.

I absolutely can’t take a generic version of my prescription

Patented drug manufacturers usually aren’t willing to sacrifice their entire market share to generic competitors. Some will start producing “generic” versions of their own drug, selling them to a partner or subsidiary who will sell them as an “ultrageneric”. So in many cases, at least one of the generic versions that’s marketed will be truly identical (except perhaps for markings) to the branded products. Now if you think this means people won’t occasionally identify the switch as the source of efficacy issues or side effects, my anecdotal experience says otherwise.

I had a negative reaction (e.g., more side effects) to the generic

There are no requirements for generic drugs to contain the same non-medicinal ingredients as the brand-name drug, so allergic reactions are not impossible.  Most of the excipients are generally inert but there have been suggestions that serious allergies may rarely affect some patients. However, complaints about intolerance in the absence of a true allergic response or documented allergy could also be nocebo effects, negative symptoms elicited by negative expectations from the patient.

Bioequivalence testing isn’t an accepted evaluation method

It’s not just generic companies that must do bioequivalence studies. The company that owns the patent may need to do the same studies. Any changes to the dosage form either during development, or after marketing the product, may necessitate bioequivalence studies to demonstrate that changes do not meaningfully alter the concentration-time curve. So if the clinical trial was with a capsule, and the final marketed product is a tablet, a bioequivalence study comparing the two may be necessary. This reassures regulators that the clinical trials continue to be relevant and applicable to the new dosage form.

I don’t care what you say, the drugs are not equivalent in the real world

The Medical Letter recently summarized some studies that have evaluated the clinical effects of generic substitution:

  • A systematic review of 47 studies that compared cardiovascular drugs found no evidence that brand-name drugs were superior with respect to clinical outcomes.
  • An RCT comparing a brand antibiotic vs. a generic antibiotic found no differences in clinical effects.
  • The FDA conducted its own evaluation of proton pump inhibitors (e.g., Prilosec/Losec). All five generics met dissolution standards.
  • Levothyroxine as an API is unstable, and the FDA recently tightened manufacturing standards for all versions. There are no well-documented cases of  therapeutic inequivalence between brands deemed to be bioequivalent. Most health professionals still recommend using a consistent brand, however.
  • Retrospective case-control studies with anti-epileptic drugs have yielded some data suggestive of therapeutic inequivalence. Given the limits of these studies (patients who have switched brands may differ in some way from those that did not switch), the data are not conclusive. Clinically, most health professionals are cautious when switching to generic antiepileptic drugs.


The science of bioequivalence evaluations for generics has been in place in most countries for more that 20 years with an established track record of therapeutic equivalence. These evaluation methods have been so successful in establishing generic drug standards that they are largely consistent between all of the major drug regulators worldwide. Consumers and health professionals alike can be reassured that generic drugs approved under these regulatory frameworks are indeed bioequivalent, and therefore, interchangeable with brand name products.

Posted in: Basic Science, Pharmaceuticals

Leave a Comment (37) ↓

37 thoughts on “Generic Drugs: Are they Equivalent?

  1. Interesting that extreme cynicism about the pharmaceutical industry co-exists in our society with such common and strong belief that expensive branded medicines are somehow superior — hopefully not in the same person, but I wouldn’t bet on it.

  2. Trod says:

    Paul, it’s just that big pharma is an evil, vile thing when they’re making you do something, like take a one time inexpensive “cure” that lasts for 10-20 years, but not when you choose to pay ten times the cost for the preference of brand name to generic. ☺

    With all respect Scott, the responsibility for switches to generics is more complicated than laying it at the pharmacist’s feet. Pharmacists, in the US at least, should be encouraged to both switch their patients to generics as well as educate them as to the benefits from doing so, on the basis of providing for the patients’ best interests. More frequently than not, this is also a corporate initiative, at least at retail pharmacy.

    Sometimes the patients will have less of a choice when their Rx formulary changes to prefer the generic and exclude the brand name; this will even go so far as to switch to a biosimilar drug that will soon be going generic.

    On the patient level, all they know is that the pills look different. The human body is a bizarre, arcane thing and they’ve relied on that one drug for so long, how can they change it now? Then they hear that some people are sensitive…then lazy doctors do tons of damage by blithely writing a prescription for their old medication after hearing their complaint, possibly even noting their “sensitivity” and a vicious cycle is born.

    Yes, this is even when dealing with “ultragenerics.” Even one really bizarre case where there was a shortage of the generic drug, the manufacturer approved filling generic prescriptions with brand-name stock. Yet people still proclaimed heightened sensitivity and all that.

    Much of this stems both from the overly aggressive ad campaigns for branded drugs as well as the general idea (in the US) that learning more about your own health somehow isn’t your responsibility, that patients are customers, and that the customer is always right.

  3. BobbyG says:

    Excellent. This post should get exposure far beyond the SBM choir out to the multi-meds population (e.g., the seniors cohort, or other chronic condition pts on Rx). I will linking it on social media and elsewhere. Very nice.

  4. windriven says:

    I typically opt for generics where available. The only instance of subjective non-equivalence I’ve noticed is with omeprazole. The Prilosec branded version invariably works flawlessly for me. But I’ve had to abort several attempts to switch to Costco’s house-branded version because it didn’t provide 24 hour relief.

    Yes Scott, there is a question buried in there: how do the requirements for time release agents or processes differ from those of the API?

  5. weing says:

    Your article raised a lot of questions for me. How do pharmacies get their generics? Is there a list of tested generics, and their manufacturers, that pharmacies are allowed to buy from? How is this testing verified? Are they able to buy from distributors that get their generics from China or India? Has testing been done on those?

    1. Scott Gavura says:

      @weing I don’t know what country you’re in, but here are some general responses. Your country/state/province may do things slightly differently:
      – pharmacies buy their generics direct or from approved wholesalers. The only generics that pharmacies can dispense are those that are approved for sale by that country’s regulator. The only generic drugs that can be interchanged automatically are those that have been deemed bioequivalent and interchangeable by that country’s regulator. Not all countries allow automatic substitution of generics.
      – Jurisdictions maintain their own list of generic drugs deemed bioequivalent and interchangeable
      – Testing is conducted by the generic manufacturer (or a third party) and those tests are reviewed by the regulator. The approval of a generic drug for sale is based on the bioequivalence data submitted.
      – The actual site of fabrication does not matter as long as the site meet’s that country’s drug manufacturing and quality standards (i.e., GMP). The supply chain for the ingredients in any dosage form may span several countries, as I described in my post on drug shortages.

  6. qetzal says:


    To the best of my knowledge, all generics sold in the US still have to get FDA approval, which requires submitting the data that proves the generics are, in fact, bioequivalent, and are manufactured according to cGMP standards. You can find out what generics are approved for any given drug at the Drugs@FDA website. I don’t think it matters whether they’re made in the US vs. China or India. Only whether they’re FDA approved.

    However, I am not a regulatory professional or FDA expert, so don’t take that as gospel.

  7. WilliamLawrenceUtridge says:

    One interesting and horrible wrinkle in the generic-versus-name brand debate is the impact on the pharmaceutical industries to do research. More profits = more money for the extraordinarily expensive task of bringing a new drug to market. Generics significantly undercut those profits and I’m guessing don’t result in many new drugs being made.

    Of course, if we did disallow generics and insisted on name brand-only, they’d probably spend the resulting profits on advertising :(

  8. TsuDhoNimh says:

    Generic drugs are manufactured differently from branded drugs, and branded drug manufacturers use better processes” … ROFLMAO! I can think of several drugs where the branded and the generic roll off the same manufacturing line and leave by the same loading dock. The only difference is the color of the shell and the ID code, and the label on the bottle.

    “I had a negative reaction (e.g., more side effects) to the generic” I have had this happen because of a substance used in the filler in the generic. Changing to a different maker’s generic drug solved it.

  9. bluedevilRA says:

    I have heard several doctors say that for medications that treat serious conditions (seizure meds) or that are potentially dangerous (coumadin) it is better to go with brand name for consistency. As in, if the pharmacy switches between generic manufacturers and there’s minor fluctuations in the drug then this could pose a serious problem for patients. Does that make sense?

  10. Trod says:

    bluedevilRA, your doctor is only half right and barking up the wrong tree for some of this. The science says there should be a relatively low percentage of people with abnormal sensitivity who would require a brand name (or different generic) — these drugs API are supposed to be verified as bioequivalent, and the non-active ingredients should have a nominal effect, if any, on this.

    Yet both classes of drugs you listed are recommended to be used with heavy physician oversight and once a physician, out of fear or whatever, begins telling horror stories about how much they desperately need to be on the brand name then you’ve got a solid case for the nocebo effect.

    Of course, most psych meds are temperamental in the first place which can lend credence to this.

    WilliamLawrenceUtridge > Rx companies already spend greater than 50% of their profits on advertising, even as they proclaim the need to keep patents and have expensive branded drugs to fund research. Some common theorycraft states that they’d probably be fine if forced to charge less or only make “generics”, being that they could easily spend less on advertising and still make the money back for research purposes.

  11. Kultakutri says:

    If I may chime in as a humble patient, I like to keep my meds consistent. It happened to me twice that for some reason, I got a different generic citalopram than the usual one and I kept forgetting taking it because the package and the pills just looked different and in the morning, before the second coffee, my autopilot would ignore round pills because the right ones are oval and ovoid.

    Okay, I’m not an idjit so I found a way how to manage by carrying the pills in my bag and getting them at work with the second coffee of the day. I however think that this aspect of getting used to new package, new shape, colour, whatevs with every prescription may lead to worse outcome.

  12. Oh shoot, Who’s the Italian dude that’s doing all the research on placebo effects (I think he’s a neurologist). Didn’t he have some research about conditioned pysiological responses to recognized pill shape and color?

    If I could remember his name, I could look it up.

  13. Harriet Hall says:

    The Italian dude is Benedetti.

    Big pills>small
    The more doses a day the better
    The more expensive the better
    Color makes a difference

  14. gippgig says:

    Speaking of “The more doses a day the better”, the blood plasma level graph brings up an unrelated point I’ve wondered about for a while – the likely advantage of staggering doses. For example, instead of taking 2 pills every 4 hours, take 1 pill every 2 hours. This should result in significantly reduced variation in the level of the API, greatly lessening the chance of dropping below the MEC or exceeding the MTC (& might in some cases allow a higher dose for greater efficacy or a lower dose for reduced chance of side effects). Has this been studied (other than as a placebo effect)?

  15. Thanks so much Harriet, my idea was that when one changes the color or shape of the pill, it might undermine the conditioned response that the patient has developed taking their name brand medication. Without the conditioned response the patient may get a smaller physiological response from the different looking medication.

    (Meaning Pavlov can’t change to a piano instead of a bell without expecting the dog to salivate less)

    Is that possible?

  16. Sorry, didn’t mean to be overly familar in addressing, Harriet Hall, that is.

    1. Harriet Hall says:

      I don’t mind you calling me Harriet. You’ve been posting for a long time and have earned my respect: I consider you a friend worthy of first-name basis. I only object to being informally addressed (particularly as”Ms. Hall”) when it appears to be intentionally disrespectful. And yes, changing the appearance of the pill would interfere with the conditioned response.

  17. CarlRichard says:

    @ bluedevilRA and @ Trod
    There is more to this The rationale in the case of antiepileptic drugs and coumadines is that plasma levels of these drugs matter a lot. Fluctuation in the plasma levels of antiepileptic drugs might lead to seizures for example. Even relatively small differences in Bioavailability between Drug A and B might matter. Therefore it is sometimes adviced to take only the drugs from one manufacturer of a given drug, but it does’t matter whether this is a generic drug or a brand drug.

  18. lilady says:

    There is a difference in brand name versus generic anticonvulsant drugs. My son was on Dilantin Infatabs, not Dilantin Kapseals which are metabolized differently…and he certainly was never prescribed generic phenytoin. He was also on Mysoline not the generic primidone. It was difficult enough, sometimes impossible, to keep his grand mal seizures under control and, I always provided his brand name medication for hospital stays. Most hospitals do not have the brand name anticonvulsants on their formularies. His brand name anticonvulsants were two of the exception medications that were covered by our drug plan at no additional cost. It is my understanding that most drug plans also exempt most of the brand name anticonvulsant drugs from their “generics only” policy.

    My husband was on Coumadin for months after undergoing separate right and left atrial ablations. The doctor specified the brand name and again he was covered under our drug plan with no additional cost.

    I’ve been on Lipitor for years. I was in one of the early Lipitor blinded trials because of genetic high lipids. I’ve have absolutely benefited by Lipitor and was notified several months ago that as of January 2012, I would be switched to the generic formulation.

    I have been reading on the web that Pfizer has been notifying the large drug plans that they will supply Lipitor for the generic price…so who knows if I will be switched when it is time to re-order my 3-months supply? I also have a rather large “stash” of the drug, so I won’t be ordering until March or April. I know intellectually that the genetic is probably the bio-equivalent of Lipitor…but ~ 1-2 months after starting the generic, I’ll have a lipid profile done.

  19. lilady says:

    oops brain fog: “I know intellectually that the genetic is probably the bio-equivalent of Lipitor…but ~ 1-2 months after starting the generic, I’ll have a lipid profile done.” “generic” not “genetic”…doh.

  20. @Harriet Hall – Thanks!

  21. Troyota says:

    I’m a longtime lurker and finally signed up so that I can say, THANK YOU for this excellent discussion of bioequivalence! The graphics are superb–the first graph is a PK lesson in itself, and a beautiful example of “one picture is worth a thousand words”. This is an A+ piece of scientific communication. Many thanks!

  22. weing says:


    I’m in Connecticut. Thanks for your reply. It answered a lot of my concerns. I recall that about 20 years ago when dyazide went generic, a generic manufacturer was found to have submitted the branded medication inside its own capsules for bioequivalence testing.

  23. Nakamova says:

    I am still a bit confused about which of the miniscule variations between brand and generic may be affecting those who take anti-seizure medicines. I have epilepsy, and I have had zero problem with switching to the generics. However, I’m familiar with quite a few folks who HAVE had seizures after being switched to the generics. They had seizure control on the brand, and lost seizure control on the generic. These are not cases where the nocebo effect is in play, since these people were unaware of any risks when making the switch. Scott, can you give me a sense of what the reasons might be?

  24. evilrobotxoxo says:

    Thanks for the excellent explanation of bioequivalence. I’m one of the commenters who raised the issue of bioequivalence in prior discussions. In general, I agree 100% with the notion that bioequivalence means clinical equivalence, and I agree that the majority of complaints about generic meds are unfounded. However, I think that the conclusion of this post is a bit oversimplistic because obviously bioequivalence equals clinical equivalence, but the real issue is whether or not the current FDA standards actually enforce true bioequivalence. For example:

    Does a slight difference in the rate or extent of absorption make a clinical difference?  Most regulators worldwide have decided that a 20% variation is generally not clinically significant.

    What’s magic about this number 20%, and what evidence is there that 20% variation in a given pharmacokinetic parameter is clinically insignificant? For the vast majority of meds it probably is insignificant, but it’s hard to argue that’s the case for all of them. For example, there are plenty of times I’ve changed the dose of a patient’s antipsychotics by less than 20% and seen significant changes in either side effects or therapeutic effect. Or would anyone argue that 20% changes in coumadin dose are clinically insignificant? Obviously, this is less of an issue that it could be because the vast majority of meds differ by much less than the maximum allowed by the FDA. But what about ones like levothyroxine? Here is an excerpt from the FDA web page linked to in the original post:

    First, the clinical experts in endocrinology on the EMDAC unanimously agreed that allowing the usual 10 percent loss during a product’s shelf life could have clinically relevant consequences. Second, the vast majority of the joint committee agreed that the potency specification for levothyroxine sodium products should be narrowed from 90-110 percent to 95-105 percent.

    The FDA itself concluded that their old bioequivalence standards did not reflect true bioequivalence in the context of levothyroxine, so they made new ones. There aren’t any high quality, large-scale RCTs to guide us here, but if an FDA panel of endocrinologists thinks that the old FDA standards for bioequivalence weren’t sufficient, I’m inclined to believe them. If even the FDA themselves acknowledge that this is true for levothyroxine, how do we know it’s not true for a small number of other medications as well? For example, every epileptologist I’ve ever discussed the issue with believes that the current FDA standards for bioequivalence with certain antiepileptic drugs is not sufficiently stringent. As pointed out in the original post, the data is inconclusive, but this is another case where I’m inclined to defer to expert opinion, partially because I think prior plausibility is on the side of the epileptologists being correct.

    A second, unrelated question for Scott Gavura: how does the FDA determine bioequivalence for time release formulations, particularly complex ones like Concerta, which has triphasic 12-hour release? Clinical lore has it that FDA bioequivalence standards break down with some of the newer generic time release formulations. I have no idea whether that’s true, and I’m not claiming that it is, but obviously one would have to use a different pharmacokinetic model for time-release vs. immediate-release formulations. Is there a standard methodology for that?

    1. Scott Gavura says:

      @evilrobotxoxo and others:

      I clearly didn’t devote enough space to pharmaceutical standards, which seem to be the source of many questions.

      Before we undertake the bioequivalence studies, there must be evidence collected to verify that the products are pharmaceutically equivalent. That is, each dosage form contains the same amount of the API, and meets the same standards for strength, quality, purity, content uniformity and dissolution times. Not only must these parameters be comparable between brands, they must be consistently comparable between lots of the same brand.

      Potency is a quality specification, not something that’s evaluated in the bioequivalence tests I have described. However, it is evaluated in making the overall determination of whether products can be substituted – that is, it’s part of the quality standard that all products must meet.

      The potency specification for levothyroxine was narrowed to eliminate the chance of any reduced-potency products being dispensed, say, near the expiry date. The specifications are applicable to all products – the original brand, and the generics. Importantly, it reduces the risk of any variation not only between brands, but between lots of the same brand. From what I can see there was no change in the bioequivalence standards for levothyroxine.

      I may not have been clear in my description of the tests: The geometric mean ratios (AUC test / AUC reference) as well as the projected 90% confidence intervals for the population mean ratio must fit completely with an 80%-125% window. If the estimates value varied in any substantial way from each other, it’s unlikely the 90% CI’s could fit within the 80% to 125% window. For drugs deemed “narrow therapeutic index”, it’s a 95% CI that must fit – even wider, meaning the curves must be nearly superimposable.

      There are a few questions in the comments on modified release dosage forms and I’ll come back to this.

  25. Scott says:

    evilrobotxoxo makes a good point about the 20%. That’s similar to the steps between dosages of levothyroxine, for instance. 100 mcg vs. 125 mcg has a pretty substantial impact, so a 20% difference in absorption can’t be credibly claimed as clinically insignificant.

  26. Scott says:

    Ah, so the 20% figure is just the default and is adjusted for the particular medication in question if needed. That makes a lot more sense.

  27. Scott says:

    A third read indicates that the adjustment isn’t in the 20% window, but I think I get it now and your explanation does make it a lot more clear – thanks.

  28. evilrobotxoxo says:

    @Scott Gavura:

    That makes sense that the levothyroxine thing was a potency issue and not a bioequivalence issue – my mistake for conflating them.

    Nevertheless, I still disagree with your assertion that bioequivalence parameters being within a relatively large arbitrarily-determined range is a meaningful guarantee of clinical equivalence. I’m not 100% sure I’m interpreting your explanation of the statistics correctly, so please correct me if I’m wrong. Basically, I think a simplified but accurate way of describing the standards would be to say that if a patient were switched from the brand name to a generic that met the bare minimum standards for bioequivalence testing, there’s a 10% chance that the total amount of medication absorbed (or the peak blood concentration) would be either less than 80% or more than 125% of what it was on brand name medication. How is that a meaningful guarantee of clinical equivalence?

  29. Quill says:

    For me this is a most helpful post that clearly explains a very important part of chosing medications. My thanks to Scott Gavura for the original post and to everyone for the comments so far.

  30. kdv says:

    An issue for me when I was in general practice was the fact that I had a very high proportion of elderly patients. For some ( by no means all ) of those, their daily medication routine was ruled by the size, shape and colour of the tablet and/or the pack in which it was supplied.

    I faced a few cases where people were found to be doubling, or even tripling their prescribed doses, because they’d obtained a supply at a different pharmacy who had a different “preferred” manufacturer.

    And yes, I did tell them to look at the generic names. Yes, I did write them down. It still happened. Fortunately, no seriously bad consequences eventuated insofar as I’m aware, but that was undoubtedly due to luck.

    There may well be a reason to switch to a cheaper alternative once a patent expires, but I can’t see any need whatsoever to be allowed to switch back and forth at the whim ( or financial interest ) of the pharmacist thereafter, as can happen here in Australia at least ( I have no idea of the rules in the USA ).

  31. weing says:


    Same problem in the US.

  32. evilrobotxoxo says:

    @weing, kdv:

    I’ve definitely had that problem, but in slightly different form with low-functioning psych patients, a lot of whom accuse you of poisoning them, refuse to take the “new” medication and decompensate as a result, etc. A semi-related problem is when insurance refuses to pay for 30 5-mg tabs, saying that I’m supposed to prescribe 15 10-mg tabs and the pt is supposed to cut them in half, or some nonsense like that. I remember one of my patients whose family wheeled him in to a followup appt in a wheelchair, completely parkinsonized, because his daughter-in-law was accidentally giving him 3X the prescribed dose of an antipsychotic because of all that pill-splitting nonsense.

    99% of the time, I’m all for generic meds and substitutions, and it’s fine with me if it works that way by default, but prescribers should have the option to exert more control over the process without having to resort to using overpriced brand-name meds.

  33. lilady says:

    I am wondering if anyone has experience with going off a hypertension medicine that worked superbly for 10 plus years ( Verapamil SR plus a potassium sparing diuretic)…then failed to adequately control hypertension…and then giving this regimen another “chance” after a few years.

    For a variety of reasons I cannot take ACE inhibitors and Ace Blockers and have been treated with Amlodipine 5mg., Chlorthalidone 25 mg. and Klor-Con 10 meq. for the past 2 years, strictly monitoring my own blood pressure.

    Would returning to the Verapamil and any diuretic (potassium sparing or non) be a valid option, should my present hypertension medicine fail to control my hypertension? (I like to have a “fall back” option/plan B in place.)

  34. mimieliza says:

    As a patient (well, actually a patient’s mother – my toddler has GERD), I have found that the issue is not generic vs. name brand, but rather, consistency. My son was on lansoprazole (Prevacid) solutabs for nearly a year when suddenly the generics were in short supply. Our pharmacy substituted the name brand Prevacid (at significant cost to me). Within a week, my son’s symptoms, which had been well-controlled, flared terribly. His pediatrician doubled the dose and the symptoms were back under control. However, now I was buying twice as much medicine at a premium price. The expense was significant, and I was also worried about putting my little one on such a high dose.

    I found a pharmacy that did not seem to have supply problems with the generic lansoprazole. We switched to that pharmacy, and my son was able to go back on his original, lower dose of the generic. Now I verify when I fill the prescription that we are receiving the generic made by Sandoz.

    From my limited understanding, PPIs can be tricky in dosing/absorption for children. For whatever reason, the way the generic is formulated works well for my son; the name brand did not work quite as well and required a higher dose to receive the same benefit.

Comments are closed.