Lot to Lot Validation

Lot to Lot Validation
Jun 29, 2019 10:33am

From Maureen (Mo) Tacke, St Luke's Health System:

Hi George, our institution is coming to the joyous time of lot roll over in coagulation, and it’s got me wondering... Is our process a little (or a lot) overkill? Current process:

  • Normal ref range studies (20 normals) for PT , PTT , FIB, AT , TT , PC , PS , DRVVT , SCT, APCR , factor VIII, factor IX, VWF activity, VWF antigen
  • Lot to Lot studies (20, spanning the range) for all tests listed above plus D-dimer.
  • Control Studies (assayed controls), 10 runs over a minimum of 5 days.

I started questioning why normal reference range studies on any assay other than the CAP mandated PT and PTT ? It is possible to shift reference ranges in the other assays, but in my experience, it’s unlikely. If we did happen to see a shift in cut-offs for ratios or reference ranges, doing a full blown study of 150 patients seems astronomicaland finding the perfect reference to verify to challenging in a short amount of time. Are 20 lot to lot specimens necessary? According to CLSI EP26-A running 2-5 specimens that span the range and hit clinical decision making cut-offs is best practice. The control studies seems reasonable as we’re verifying precision of the new lot. I guess I’m asking, is all of this necessary, and if so, what reference indicates this? I’m considering our new process as follows:

  • Normal ref range studies (20 screened normal patients) for PT and PTT
  • Lot to lot studies (2–5 specimens), spanning the range and hitting clinical decision making cut-offs--must be within TAE of 20%.
  • Control studies (assayed controls), 10 runs over a minimum of 5 days. Compare CVs to stated IFU or historic runs.

Thanks in advance for enlightening, Mo!


George responds: Hi, Mo, thanks for your question. I've been attending the ASCLS meeting in Charlotte, NC, giving me the opportunity to discuss the question with some colleagues. You are right, you can rely on your lot-to-lot studies to determine whether you need to re-compute your reference intervals. If the lot-to-lot results are within your parameters, you can continue to use your existing reference intervalss as you outlined. Many of us use five specimens that represent the full reference interval, or even use two, one high and one low and mix them to prepare intermediate levels. Of course, if you are using the PTT to monitor heparin, you will have to perform the dreaded "Brill-Edwards" curve, comparing PTT and anti-Xa results. The target values for PTT reagent always vary from lot to lot.

If you pardon the self-reference, the requirements and methods are provided in DeVries H, Fritsma GA. Chapter 2, Quality Assurance in Hematology and Hemostasis Testing. In Keohane EM, Otto CN, Walenga JM. Rodak's Hematology; Clinical Principles and Applications, Sixth Edition, Elsevier Press, 2019. The chapter is fully referenced to primary documents.

0 Comments

From Maureen (Mo) Tacke, St Luke's Health System:

Hi George, our institution is coming to the joyous time of lot roll over in coagulation, and it’s got me wondering... Is our process a little (or a lot) overkill? Current process:

  • Normal ref range studies (20 normals) for PT , PTT , FIB, AT , TT , PC , PS , DRVVT , SCT, APCR , factor VIII, factor IX, VWF activity, VWF antigen
  • Lot to Lot studies (20, spanning the range) for all tests listed above plus D-dimer.
  • Control Studies (assayed controls), 10 runs over a minimum of 5 days.

I started questioning why normal reference range studies on any assay other than the CAP mandated PT and PTT ? It is possible to shift reference ranges in the other assays, but in my experience, it’s unlikely. If we did happen to see a shift in cut-offs for ratios or reference ranges, doing a full blown study of 150 patients seems astronomicaland finding the perfect reference to verify to challenging in a short amount of time. Are 20 lot to lot specimens necessary? According to CLSI EP26-A running 2-5 specimens that span the range and hit clinical decision making cut-offs is best practice. The control studies seems reasonable as we’re verifying precision of the new lot. I guess I’m asking, is all of this necessary, and if so, what reference indicates this? I’m considering our new process as follows:

  • Normal ref range studies (20 screened normal patients) for PT and PTT
  • Lot to lot studies (2–5 specimens), spanning the range and hitting clinical decision making cut-offs--must be within TAE of 20%.
  • Control studies (assayed controls), 10 runs over a minimum of 5 days. Compare CVs to stated IFU or historic runs.

Thanks in advance for enlightening, Mo!


George responds: Hi, Mo, thanks for your question. I've been attending the ASCLS meeting in Charlotte, NC, giving me the opportunity to discuss the question with some colleagues. You are right, you can rely on your lot-to-lot studies to determine whether you need to re-compute your reference intervals. If the lot-to-lot results are within your parameters, you can continue to use your existing reference intervalss as you outlined. Many of us use five specimens that represent the full reference interval, or even use two, one high and one low and mix them to prepare intermediate levels. Of course, if you are using the PTT to monitor heparin, you will have to perform the dreaded "Brill-Edwards" curve, comparing PTT and anti-Xa results. The target values for PTT reagent always vary from lot to lot.

If you pardon the self-reference, the requirements and methods are provided in DeVries H, Fritsma GA. Chapter 2, Quality Assurance in Hematology and Hemostasis Testing. In Keohane EM, Otto CN, Walenga JM. Rodak's Hematology; Clinical Principles and Applications, Sixth Edition, Elsevier Press, 2019. The chapter is fully referenced to primary documents.

Leave A Comment

You must be logged in to Comment - Sign In