I am working developing/validating an impurity HPLC method. It’s a pretty straightforward HPLC method, with a main peak (monomer) and an impurity peak (dimer).
To evaluate linearity/accuracy/range per ICH guidelines, the classic approach is to take an ultrapure sample and to spike with known amounts of the dimer or with known amounts of a sample with high dimer content. The sample concentration should remain relatively constant (i.e. monomer peak area is constant). The spike recoveries can then be used to evaluate method linearity/accuracy.
Another approach I’ve heard used in method development is to take a sample with high dimer content (e.g. 2%) and serially dilute this sample until you no longer get accurate + precise dimer quantification. For example, a 1:20 dilution may be the most you can dilute this sample down while still getting accurate + precise dimer quantification. You could then divide 2% by the dilution factor to get the lower end of your assay range (i.e. LLOQ; 2%/20 = 0.1 %).
I’m not used to the second approach (the serial dilution). Can anyone tell me what I might be overlooking with this approach? I know it doesn’t follow ICH guidelines, but from a scientific point of view, why or why is it not appropriate?
Hello, serial dilution is widely used in HPLC method development and validation. Only thing use the rage with is corresponds the range of analyte in future analyzing samples.
Dear Sam
As per ICH guidelines if you go for serial dilution there may be less chance of manual error and if you considering the Linearity you will get good regression in serial dilution even your working with wide linearity range (1000ug to 10ng).
Hello, serial dilution is widely used in HPLC method development and validation. Only thing use the rage with is corresponds the range of analyte in future analyzing samples.
If I have 2% dimer and I serially dilute the sample down to a 1:20 dilution and I still have 2% dimer. Is it then appropriate to conclude that my method is capable of quantifying a sample with 0.1 % dimer?
Serial dilutions are widely used by chemists. Two common mistakes: (1) Volume based dilution is far less precise and less accurate than weight based dilution. The dV of mixing and density effects are far less than the inaccuracy from using volume. The most serious errors arise from temperature, evaporation, wetting and contamination.(2) A single dilution series is used, rather than two dilution series starting from the same master stock. Comparing the curves from the two series will provide much more robust results. The results will reveal imprecisions in the technique quite nicely.
Yes, if % concentration is a factor of purity and accuracy using CoVo = CrVr.
Where C - concentration, V = volume
o = original, r = required
Then plot a standard curve graph and extrapolate.
Dear Sam Mallonee,
First approach is correct, since we don't know (at this stage) matrix interference due to either monomer or diamer.
First approach should be used over second. In second case, by diluting sample you'd also dilute main peak. Validation requires accurate quantification of dimer over the range in presence of same concentration of monomer.
Generally, when you want to dilute a sample a get a lower concentration, serial dilution approach is the best one. You can prepare a stock solution having a higher concentration (which is easy to prepare) and then carry out serial dilutions using bulb pipettes. For each dilution, you can take a higher volume and then dilute it. This will automatically reduce pipetting error and the accuracy of the results will be good. Serial dilution is very popular in many techniques like HPLC, AAS, UV etc.
There are a few things to keep in mind when making serial dilutions. The uncertainty of the standard concentrations should be maintained at a lower level than the expected uncertainty in the measured quantities. Replicating the results is important, so making two or more separate base standard solutions and replicating the serial dilutions will give you additional information about the actual uncertainty in the standard solutions. Avoid using the same pipets when making the serial dilutions and rinse the pipets a couple of times with the standard before pipetting the standard to address any loss due to adsorption in the pipet. As you go through the series of dilutions, the uncertainty of the standards will add up. As far as doing gravimetric dilutions, as all HPLC measurements are strictly volume measurements, the potential gains from doing gravimetric dilutions are then convoluted with uncertainties in solution density based upon potential volume non-additivity, temperature variations resulting in density fluctuations and evaporation losses. Sometimes these things matter and sometimes they don't.
Deluting your sample means that you are deleting your matrix. Therefore the matrix effect/interferance with your analyte will be NOT taken into account in your calibration curve ..
Serial dilution May not be precise for quantifying impurities
The first approach with spiking your sample with the expected range of concentrations is more accurate
Sam, serial dilution is a good way to determine how your HPLC test method is behaving from a methods development perspective. Once you've determined the LLOQ for the dimer impurity (concentration, in ug/mL, for example), you'll need to indicate if it is at an appropriate level in regards to the monomer (% impurity, for a quantitative or limit test method analysis). Does it meet your requirements?
I agree with all the earlier answers to your original question. From a validation perspective, the ICH guidelines are pretty clear: the evaluation of the dimer impurity must be performed in the same test matrix that would typically be found under normal sample preparation conditions. This can be performed artificially using concentrated stock solutions of the monomer and dimer, and then diluting them to the required final concentrations. You can "wash out" the effect of the sample test matrix if you do not maintain the correct concentration of the monomer (i.e., it is not a true representation of the impurity profile in the presence of the monomer).
For example, the performance characteristics of accuracy can be performed via spiked recovery or sample addition studies. If the dimer can be provided separately with a known independently derived potency/purity level, then the accuracy evaluation is fairly straight-forward.
- Badges
- Science method