Bioequivalence and Bioavailability Forum • compromise

compromise [Bioanalytics]

posted by Helmut – Vienna, Austria, 2014-06-19 03:13 (3994 d 03:45 ago) – Posting: # 13089
Views: 14,360

Dear Ohlbe,

❝ The reason is usually something like […]

I’m afraid you are very right!

❝ The 50 % in the EMA guideline is probably too high, but with a log placement the MQC gets rather low, and leaves a huge gap between MQC and HQC. With a log placement of calibrators, there is also a huge gap between the HLOQ sample and the one before.

The gap disappears if you start thinking in log-scale (which makes sense, IMHO). Let’s explore the example from above (x = 2–200, n = 10). I would place the LQC at St2 (5.57 = 2.78 × LLOQ) and the HQC maybe between St9 & 10 (170 = 85% of ULOQ). How do two MQCs behave? MQC₁ based on the arithmetic mean (101, placed between St8 & 9) and MQC₂ based on the geometric mean (20, between St5 & 6). Whereas MQC₂ splits the curve in halves (five calibrators below/above), MQC₁ is extremely asymmetric (eight below, two above).

❝ It may not be a problem with a linear response, but with a quadratic equation this gap between calibration samples does not help to fit the curve…

There is no gap in calibrators.

❝ (where does it become non-linear ?)

Nowhere; it’s nonlinear in the entire range. Commonly the coefficient of the quadratic term is negative – with increasing concentrations the local slope decreases.

❝ and the QCs may not help to see inaccuracies.

Have to think about it.

❝ Personally I would set the MQC around 1/3 of the ULOQ,…

In my example that would still be a 7/3-split of the curve.

❝ …but I have strictly no data or rationale to defend it.

❝ In any case there may be a need to adjust the concentration of the QC samples based on the concentrations found in the subject samples after the first few runs, if the calibration range is not optimal.

Good point!

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes

Complete thread:

determining the QC medium haydonat 2012-10-18 00:26
- QC medium: short answer Helmut 2012-10-18 16:14
  - QC medium: short answer ElMaestro 2012-10-18 16:51
- QC medium: lengthy answer Helmut 2012-10-19 15:47
  - QC medium concentration in 2014 Debbie 2014-06-10 19:42
    - ~geometric mean of range Helmut 2014-06-17 14:30
      - ~geometric mean of range nobody 2014-06-17 14:49
        
        ~geometric mean of range Helmut 2014-06-17 15:11
        
        ~geometric mean of range nobody 2014-06-17 18:26
        
        ~geometric mean of range Helmut 2014-06-18 16:07
        
        compromise ElMaestro 2014-06-18 17:43
        
        compromise Ohlbe 2014-06-18 20:00
        
        compromiseHelmut 2014-06-19 01:13
        
        compromise Ohlbe 2014-06-19 11:15
        
        compromise nobody 2014-06-19 14:10