Pathology Laboratory Testing

Front 1

. How would you establish a reference interval for a diagnostic test? Mention how you would obtain the data, and how you might analyse them.

Reference intervals

Back 1

usually include the central 95% of the distribution (i.e. mean  2x SD). i.e. a Gaussian distribution (tho’ not all distributions are Gaussian tf. Non-parametric statistics are sometimes used).

Front 2

IFCC reference values definitions:

Back 2

Reference individuals comprise a reference population providing a reference sample group, which gives reference values, which provide a reference distribution, to calculate reference limits, which may define the reference intervals.

Front 3

Establishing a reference interval:

Back 3

1. Define the reference population
2. Sample randomly from it
3. Assay the variable in question
4. Study the distribution of the data
5. Decide on the appropriate calculation
6. Calculate the reference limits and their confidence intervals

Front 4

Reference intervals from a total patient population:

Back 4

Use a method to define the probable ‘normal’ population; it is usual to put arbitrary limits on the population data to remove gross outliers.

Evaluate on probability paper or with a program for the Bhattacharya method – to confirm intervals are relevant to the population.

Front 5

Intra-individual variation and reference intervals:

Index of individuality

Back 5

Index of individuality (= CVi / CVg):
 If <0.6: the variable is said to have high individuality and conventional reference intervals will not be of much use (e.g. with alkaline phosphatase).
 If >1.4: the reference interval will be useful (e.g. with calcium).

Front 6

Intra-individual variation and reference intervals:

Index of heterogeneity

Back 6

The ratio of the observed CV of a set of individual variances (including CVa) to the theoretical CV [2/ (n-1)], where ‘n’ is the mean number of samples collected per subject. The SD of the difference between this and unity is [1/ (2n)]. If the ratio differs by at least twice this SD there is significant heterogeneity.

Front 7

What are the likely effects on laboratory test results of the following?

i. EDTA contamination

Back 7

EDTA chelates calcium: inc. K and dec. levels of Ca, Mg, and ALP.

Front 8

What are the likely effects on laboratory test results of the following?

ii. Haemolysis

Back 8

Haemolysis contaminates blood tests with RBC constituents…
May be due to:
 Skin cleansing agent
 Blood withdrawal by forceful suction
 Transferring blood from syringe to container through needle
 Inappropriate amount of anticoagulant/preservative
 Vigorous shaking of blood
 Prolonged storage or refrigeration
Results in:
 Significant  serum levels of LDH, AST, K and ALT (relative ratio - 160:40:23:7 respectively in RBC compared to: 1 in serum)

Front 9

What are the likely effects on laboratory test results of the following?

iii. Non-specific illness

Back 9

Non-specific illness may alter testing as the body reacts:
 Acute-phase proteins
 dec. Serum Fe concentration
 Glucose intolerance
 Non thyroidal illness
 dec. Serum cholesterol

Front 10

What are the likely effects on laboratory test results of the following?

iv. Posture

Back 10

Some instances it may be important to use standardized posture:
 Upright posture  10-20% hemoconcentration (fluid leak to extravascular space)
 Standing or sitting Vs recumbence:  up to 20% of plasma proteins and protein bound metabolites (e.g. cortisol, T4, Ca and Fe).

Front 11

What are the likely effects on laboratory test results of the following?

v. Venepuncture from a drip arm

Back 11

Likely effects of a drip include:
 Dilutional effect on blood
 Saline drip: inc. Cl, dec. all other constituents (except Na)
 Dextrose drip: significantly inc. glucose, dec. all other constituents
 K in drip: inc. K