Pitfalls in Interpretation of Clinical Laboratory Tests Flashcards
What information do we need to interpret the blood sample?
-Dietary restrictions
=Fasting- glucose, cholesterol, triglycerides (overnight)
-Timing
=Diurnal variation (cortisol, testosterone in males)
=TDM (time from last dose- peak or trough)
=Stability (ATCH stable 30 minutes)
-Affected by venous stasis?
=Protein bound components (Increase C2+ when tourniquet on too long)
-Posture
=Renin and aldosterone (lying down for 15 mins before)
What tube type is required for different tests and what are the collection additives?
- Plain serum (no gel)= for clotted blood, left for 30 mins
- Plain serum (gel)= gel for barrier so that it can be spun and easily separate rec cells
- Lithium heparin (anticoagulant)= used in emergency situations so 30 mins not needed, paediatrics so more plasma obtained
- Potassium EDTA (anticoagulant)= preservative for taking full blood count, binds calcium so stops platelets clumping
- Tri-sodium citrate (anticoagulant)= binds calcium more reversibly, clotting studies (INRs), LFT
- Sodium fluoride/ potassium oxalate (anticoagulant)= binds calcium, fluoride inhibitor for respiration so stops metabolism of glucose
How do we work out the reference range?
-Reference population= population that does not have disease, consider affects of sex and age differences
-Plot reference population of at least 120 subjects (best) in normal distribution
-Mean +- SD = 95% of results
=Reference range is mid 95% of population studied
*normal range inappropriate as does not allow for statistical outliers
Describe the TSH distribution and how this affects UK guidelines
-Skewed (mean to right as more high values= long tail)
-Same rule applies for reference ranges
Treating sub-clinical hypothyroidism
=TSH >10 mU/L start on thyroxine
=TSH raised but <10 mU/L
There is no clear evidence to support the benefit of routine early treatment with T4 in non-pregnant patients
*varies greatly with equipment used to measure
What factors affect reference ranges?
-Precision (how reproducible is a result)
=poor precision broadens reference range
-Accuracy (how near the true value is the result)
=inaccurate methods moves population curve depending on bias (variation in labs)
When might we no not have a reference range?
When ‘normal’ population associated with gradient of risk
=Cholesterol
=Concept of desirable ranges
How do we decide a test result has a clinically significant change?
- Test subject to analytical variation and biological variation (within and between days, physiological variation)= TSH higher at night
- Critical Difference calculation (2.8 x sqr root of SD(A)^2 + SD(B)^2)
What is a cut-off?
- Often overlap between healthy and diseased populations
- Where values and disease meets reference range
- False negatives= with disease below reference range
- False positives= diagnosed with disease but healthy
What does inclusion of the full reference population in the diagnostic cut-off lead to?
- Good clinical specificity (few false positives)
- Poor clinical sensitivity (many false negatives)
What does inclusion of all diseased populations in the diagnostic cut-off lead to?
- Poor specificity (many false positives)
- Good sensitivity (few false negatives)
How does screening tests use specificity and sensitivity to decide on cut-off?
-Need to know prevalence of disease in a population
-Prevalence= number of people in a population with the condition
=When screening for rare disease, tests of extremely high specificity and sensitivity required
=95% cut-off not relevant for screening for rare disorders
What are the types of errors in laboratory testing?
- Pre-analytical
- Analytical
- Post-analytical
What are pre-analytical errors?
-Failure to choose the correct tests in the correct manner on the correct patient
=Spurious hyperkalemia/ hypocalcaemia- contamination from K+ EDTA anticoagulant additive
-Biochemistry draws should be done first (order of draw important)
What are analytical errors?
-Is the test appropriate for clinical need?
-Are there interfering factors?
=Interference in TSH immunoassay
=Sample treated to remove interfering heterophilic antibodies
Examples of analytical errors
- Elevated hCG= unnecessary chemotherapy, hysterectomy
- Falsely low Digoxin= overdose leading to toxicity
- Low insulin= confusion over diagnosis of insulinoma
- Falsely high troponin= incorrectly diagnosed myocardial infarction
