i have had a hard time getting the MYLAN 50/200 and ran across this report from the NPF, interesting history on MERCK transitioning sinemet manufacturing to MYLAN - need to check what the story is now - and general discussion on variability of generic quality and potency.

http://www.parkinson.org/NationalPar...22290efb60.pdf

I recently got around to reviewing this link and I thought it was important to discuss a major misunderstanding that many people have about the FDA's bioequivalency standard for generic drugs. This misunderstanding was clearly evident in the erroneous information that PDF put out in the 2011 linked report. Please keep in mind that I am in no way defending current requirements as being sufficiently rigorous or not. I will just try to explain the fallacy that people have about there being an allowance of as much as 45% difference when testing drugs for their bioquivalencies as the PDF reported.

Bioequivalency testing of drugs is usually done on healthy people, with sample sizes of less than 100. The healthy people are divided into two groups, one receiving the brand and the other receiving the generic. Blood tests are then used to determine if the drugs are absorbed and removed from the body at a similar rate. So, what similar rate would actually determine if they are bioequivalent?

This is where the FDA's 80% to 125% band comes into play, and ultimately, the misunderstanding. This band does NOT allow for a 45% variance in a generic drug's active ingredient. In actuality, it is a statistical term related to the 90% confidence interval for the "area under the curve" or AUC. This confidence interval takes into account the normal absorption and excretion differences between people. The entire AUC for the generic study group must fall entirely between the 80 to 125% band of the brand group to be considered bioequivalent. This is a much STRONGER test of equivalency than if we just compared the mean absorption and elimination amounts between the two groups because it accounts for outliers at the extremes. If fact, although not impossible, it is highly unlikely that the actual mean variance between the two groups could be greater than 5% (a typical level to determine statistical significance) and still have the AUC fall within the 80 - 125% band. Most studies have actually shown the average mean variance to be in the 3 - 4% range. In fact, this is the same range of variance normally found when testing the bioequivalency of different batches of the same drug.

Now, some can argue to tighten the band even more and this has been done for some drugs. Some "critical dose" drugs are required to have AUCs fall within a 90 - 112% band.

Once again, I am not trying to make a case for generics over brands, or vice-versa, or company vs company. I just wanted to point out that there appears to be a huge fallacy in the understanding of bioequivalency guidelines. The PDF statement that the FDA allows a generic drug to deliver between 80 and 125% of the medication of a brand is just not true. This all has to do with the normal differences among people and the old bell curve that we all remember from school. Maybe we should be more focused on these differences as they may be attributed to important factors such as weight and sex.

thanks for that clarification. that said, i have seen a number of posts where pd'ers have been forced to switch from the brand name requipXL to generic and noticed a significant difference so possibly harder to copy a CR? I have read that there might be instances where the brand name isn't manufactured anymore so the FDA compares the new generic to another generic which might have less active ingredient than the original brand name which can then result in even lower active ingredient. might just be a theoretical situation, dunno.

just had to do some googling on generics, seems FDA doesn't require the speed of release on CR drugs to be the same. might be an isolated incident but something to think about.
wonder what pharmacists say candidly when asked cuz they are swtiching generics all the time due to being forced to dispense the cheapest generics at times by insurance companies so are getting feedback all the time from patients.
they are on the front line.
http://fortune.com/2013/01/10/are-ge...branded-drugs/

also seems BRAND NAMES get taken off the market quite often when generics become available, wonder how FDA approves new generics?

just trying to be difficult.

Actually, I was just responding to your previous question and your comment above is 100% correct. Bioequivalence does not address, nor compare, time to maximum concentration. In addition, it doesn't compare anything outside of the active ingredient. Keep in mind that the majority of a "pill" is made up of "inactive" fillers which can be, and are, different from brand to brand. It is very possible the filler could interact with the active ingredient to effect absorption time. This may not be very evident in most medications (e.g. antibiotics, hypertension, gastro, cholesterol, etc). However, it could be dramatic with a PD med when timing and absorption issues are huge.

Another potential factor is that the test is usually done on only one dosage level and then assumptions are made for the other dosages. The tests are done by the company and you have to trust in their reported results, which are based on very small sample sizes (e.g. 24 - 36).

Just some thoughts.

Tupelo3, you write:

"The entire AUC for the generic study group must fall entirely between the 80 to 125% band of the brand group to be considered bioequivalent."

AUC is an area defined by the plasma concentration line and the x-axis, time. Do you mean that the 80 to 125% band refers to the plasma concentration line for all times?

You're right that comparing each point on the curve is a stronger condition than just comparing the brand AOC with the generic AOC. but it still leaves the door open to dilution: perhaps replacing 10% of the active substance with inert material would no where take the generic curve below 80% of the brand curve.

Am I correct in thinking that you refer to the pharmacokinetics of the drugs. Some drugs will not have a linear relationship between the plasma concentration and the effect the drug has on the body, the pharmacodynamics of the drug. So, for instance, a 10% decrease in concentration could lead to a 20% decrease in performance.

John

John, these are all good questions. I apologize to those that I'll bore with the answers and please feel free to skip over the remainder of this response.

1. Bioequivalence studies are cross-over studies in which
each subject acts as their own control (i.e. each gets both the brand and the generic).
2. The AUC is a measure of the ratio of the concentration of each drug over time as it is absorbed into the body for each subject
3. Bioequivalence is based on a comparison of ratios where the ratio of generic to brand does not differ by more than 8:10 (as the FDA has determined that a maximum allowance of 20%+/- is acceptable). It is based on ratios where the nominal equality is 1. It is not based on differences in absolute values or arithmetic means.

This is how the 80%-125% range for the confidence intervals of the ratio generic/brand is determined:

8/10 = 80% (lower limit)
10/8 = 125% (upper limit)

4. Using the information above, you can see how it is not possible to have a45% difference is blood concentration between the two drugs. In fact, even the maximum difference of 20% is not really possible given it could only happen if there was zero variance among the subjects on the generic test results. We all no that doesn't happen on any human test and is literally impossible in pharmacology given all of the differences among people (e.g. sex, weight, metabolism, body mass, etc.).

The 90% Confidence Interval of 0.8–1.25 reflects the limits for a comparison of ratios where equality equals 1. It is not a direct measure of the difference in systemic concentrations of the active ingredient resulting from administration of the two medicines. The confidence interval provides a range of values in which we can say with a degree of certainty the true value lies. For example, if in a study the observed ratio for maximum concentration (C max) is 0.95 (representing a 5% difference between products) and the 90% confidence interval was 0.85 to 1.01, this means that we could be confident that if the same study was conducted 100 times, then 90 of those times the observed result for the ratio of C max would lie somewhere in the range 0.85 to 1.05. This would be considered bioequivalent. Alternatively, if the 90% confidence interval was 0.75 to 1.05, this would be out of the lower range and not be considered bioequivalent.

So to summarize this ridiculously long answer, for which I apologize, in reality, for a medicine to demonstrate bioequivalence, the ratio of the mean arithmetic values (which we are more familiar with) must be close to 1 in order for the upper and lower limits to be contained within the accepted range, and any difference in bioavailability is statistically almost certain to be less than 10%, and in practicality, less than 5%. The reason is that any difference greater than 5% is likely to push the far end of the AUC either below or above the 80% - 125% band.

Regarding your second question, to the degree I understand it, the only tests conducted are serial blood plasma concentrations and urine analysis on healthy subjects. This is not an efficacy test and there is no measurement regarding performance. The drug is already approved and considered to have efficacy. I don't see that there is any importance as to whether the drug has a linear or nonlinear effect. I'm not sure I understand the last part of your question involving concentration and performance.

I do wish I was able to insert a drawing here, because one picture would have probably been able to save me many words. It's much easier to understand the bioequivalence rules when visualizing it.

Best,

Gary

Tupelo3 thanks for your reply.

For a step-by-step explanation of the process in general, not specifically FDA, see:

"Clinical Trials A Practical Guide to Design, Analysis, and Reporting"
Duolao Wang and Ameet Bakhai
http://www.richmondpharmacology.com/.../Chapter13.pdf

As I see it, there are two key questions for PwP:

Question 1: Is it possible for a generic to have substantially less active ingredient, e.g. levodopa, than the original?

Question 2: could such a difference be felt?

Suppose the nominal strength of a pill is 100mg of levodopa. It would be incorrect to say that the FDA rules imply that anything in the range 80-125 is acceptable. This would be wrong for two types of reason:
- more hurdles have to be crossed: AUC, Cmax etc.;
- the 90% confidence interval for each measure needs to be within the bounds, so the more variability in the data the wider the confidence interval and the closer to the middle you'd have to be to pass.

The danger it seems to me is that if the confidence intervals for each of the hurdles are well within the 80-125 range, it would be easy to reduce the amount of the active ingredient and still be compliant.

The rules relate to pharmacokinetic effects, but PwP are more interested in the effect the drugs have on them. This is complicated: it is not just a case of saying that a 1% reduction in the amount of a drug has a 1% effect on a PwP. Take an example: suppose a PwP takes levodopa every 4 hours, and just avoids an "off" before the next dose kicks in. A smaller dose will now create an "off" period where none existed before. This is an example of a non-linear effect.

Finally, how might pharmaceutical companies act to take into account the flexibility given by the generic rules? A generic producer might go for quality to maximise reputation, or risk reputation to save on cost. The original manufacturer while still selling the brand drug, could also make lower quality generics, while still in the acceptable range, with the intent that this muddies the waters for generics in general, thus helping to maintain the sales of the brand drug.

Thakur and Ramacha [1] describe the strategies that innovator company's can adopt with generics.

References

[1] "Pharmaceutical Business Strategy: A Generics Perspective"
Thakur and Ramacha
Journal of Intellectual Property Rights, Sept 2012
http://nopr.niscair.res.in/bitstream...%20484-496.pdf

John

title really has nothing to do with the current topic of discussion and maybe some people who want to join the discussion aren't reading this thread.