SSA 2.1 pre-analytical phase
Common errors:
1. Ordering
2. Collection – Wrong tube or not enough volume
3. Receiving
How to prevent?
Use (standardized) guidelines protocols, automation, and train personnel.
IL 2.4 Blood biomarkers - design
Sample preparation pre-analytical phase
75% of errors in pre-analytical phase
Serum = schoon, snel
Plasma = duurt lang
Why do you use tourniquet?: Loosen it, otherwise red blood cells get destroyed (hemolysis) and your
serum/plasm will turn red. increases potassium levels
Multicentre vs singlecentre
In multicentre intercenter variation can occur, due to: variation of tubes/tests, storage conditions,
transport conditions, different methods. However, multicentre gives larger heterogeneity
(generalizabilit).
IL 2.6 Blood biomarkers – design pre-analytical phase
When doing voluntary participants, allow blood drawing at regional clinicians. When doing patients
you can use drawn blood at the hospital.
When doing badges over long time: use a control that is measured doing every batch, in this case you
can see if there are differences between badges due to different reasons (or not due to the
intervention).
LE 3.1 Analytical validation
Accuracy: Describes the closeness of a attest result
obtained by the method to the true value of the analyte.
Precision: Closeness of individual measure of an analyte
when the procedure is applied repeatedly. (Inter-rater
reliability)
Expressed in (analytical) Coefficient Variation (CV%) =
relative SD = (SD/mean) x100%
LLOD: the lowest value that significantly exceeds the
measurement of a blank sample
LLOQ (quantification): The lowest concentration of a
measurand that can be determined with an acceptable
level of precision an accuracy.
, Both LLOD and LLOQ are determined using a blanc measurement (in order to see the amount of noise
that’s in here)
Sensitivity: ability of method to assess small differences in concentration of an analyte (synonym of
LLOD or LLOQ)
Specificity: ability of an assay procedure to determine concentration of the target analyte without
influence from potentially interfering substances. (how sure are we that we get signal from the
analyte we’re interested in, so adrenaline vs noradrenaline)
Sources of variation:
Preanalytical variation – during protocol different hospitals or protocols
Analytical variation (CVa) – during measurement high temperature, pipetting error
Within-subject biological variation (CVi) – variation per day or month e.g.
Between subject variation (CVg) – female/male or old/young
Between laboratory variation (Inter-rater reliability) – variation between protocols
Total allowable error (TAE%): an analytical quality specification that sets limits for imprecision AND
bias that are acceptable in a single test result. (if the absolute difference between the measured
value and the actual value is lower than the TAE the patient result is acceptable)
= z-score x0.5 CVi + 0.25 wortel(CVi2 + CVg2)
If historical values are:
- Absent: use reference ranges
- Present:
Individuality index (II) = CVi/CVg
> 1.4 use of conventional reference intervals can be used
< 0.6 better to use RCV to follow an individual
Reference Change Value (RCV): to assess whether serial results of a biomarker concentration in time
are statistically different. (So when a relative difference between serial measurements <RCV%, these
measurements do not differ)
= wortel(2) x z-score x wortel(CVi2 + CVa2)
Analytical imprecision (CV% of multiple measurement of a sample)
Desirable <0.5 CVi (so you’re looking at CVa < 0.5 CVi if yes: precision is acceptable)
Analytical bias (%)
Desirable <0.25 wortel(CVi2 + CVg2)
(Optimal < 0.125)
(Minimal < 0.375)
SSA 3.2 Analytical validation
Imagine that you aim to assess the genetic determinants of the peptide hormone hepcidin. The
hepcidin assay, however, is not yet validated. Now answer the following questions. For your answers
you may use the following information:
The hepcidin assay is worldwide not standardized or harmonized.
Your validation provided a between day imprecision of 17% at 1 nmol/L and 13% at 4 nmol/L.
The literature reports an intraindividual CV of 48.8% and interindividual CV of 154.1%.
Common errors:
1. Ordering
2. Collection – Wrong tube or not enough volume
3. Receiving
How to prevent?
Use (standardized) guidelines protocols, automation, and train personnel.
IL 2.4 Blood biomarkers - design
Sample preparation pre-analytical phase
75% of errors in pre-analytical phase
Serum = schoon, snel
Plasma = duurt lang
Why do you use tourniquet?: Loosen it, otherwise red blood cells get destroyed (hemolysis) and your
serum/plasm will turn red. increases potassium levels
Multicentre vs singlecentre
In multicentre intercenter variation can occur, due to: variation of tubes/tests, storage conditions,
transport conditions, different methods. However, multicentre gives larger heterogeneity
(generalizabilit).
IL 2.6 Blood biomarkers – design pre-analytical phase
When doing voluntary participants, allow blood drawing at regional clinicians. When doing patients
you can use drawn blood at the hospital.
When doing badges over long time: use a control that is measured doing every batch, in this case you
can see if there are differences between badges due to different reasons (or not due to the
intervention).
LE 3.1 Analytical validation
Accuracy: Describes the closeness of a attest result
obtained by the method to the true value of the analyte.
Precision: Closeness of individual measure of an analyte
when the procedure is applied repeatedly. (Inter-rater
reliability)
Expressed in (analytical) Coefficient Variation (CV%) =
relative SD = (SD/mean) x100%
LLOD: the lowest value that significantly exceeds the
measurement of a blank sample
LLOQ (quantification): The lowest concentration of a
measurand that can be determined with an acceptable
level of precision an accuracy.
, Both LLOD and LLOQ are determined using a blanc measurement (in order to see the amount of noise
that’s in here)
Sensitivity: ability of method to assess small differences in concentration of an analyte (synonym of
LLOD or LLOQ)
Specificity: ability of an assay procedure to determine concentration of the target analyte without
influence from potentially interfering substances. (how sure are we that we get signal from the
analyte we’re interested in, so adrenaline vs noradrenaline)
Sources of variation:
Preanalytical variation – during protocol different hospitals or protocols
Analytical variation (CVa) – during measurement high temperature, pipetting error
Within-subject biological variation (CVi) – variation per day or month e.g.
Between subject variation (CVg) – female/male or old/young
Between laboratory variation (Inter-rater reliability) – variation between protocols
Total allowable error (TAE%): an analytical quality specification that sets limits for imprecision AND
bias that are acceptable in a single test result. (if the absolute difference between the measured
value and the actual value is lower than the TAE the patient result is acceptable)
= z-score x0.5 CVi + 0.25 wortel(CVi2 + CVg2)
If historical values are:
- Absent: use reference ranges
- Present:
Individuality index (II) = CVi/CVg
> 1.4 use of conventional reference intervals can be used
< 0.6 better to use RCV to follow an individual
Reference Change Value (RCV): to assess whether serial results of a biomarker concentration in time
are statistically different. (So when a relative difference between serial measurements <RCV%, these
measurements do not differ)
= wortel(2) x z-score x wortel(CVi2 + CVa2)
Analytical imprecision (CV% of multiple measurement of a sample)
Desirable <0.5 CVi (so you’re looking at CVa < 0.5 CVi if yes: precision is acceptable)
Analytical bias (%)
Desirable <0.25 wortel(CVi2 + CVg2)
(Optimal < 0.125)
(Minimal < 0.375)
SSA 3.2 Analytical validation
Imagine that you aim to assess the genetic determinants of the peptide hormone hepcidin. The
hepcidin assay, however, is not yet validated. Now answer the following questions. For your answers
you may use the following information:
The hepcidin assay is worldwide not standardized or harmonized.
Your validation provided a between day imprecision of 17% at 1 nmol/L and 13% at 4 nmol/L.
The literature reports an intraindividual CV of 48.8% and interindividual CV of 154.1%.
