In many scenarios, a diagnosis is based predominantly on the patient’s history, with supporting evidence from physical
examination and laboratory investigation. Diagnosis is rarely based on the results of laboratory investigations alone,
and testing is not always the best clinical course of action. Consideration needs to be given to whether testing will
add meaningful information to the overall clinical picture, and then to the factors which may need to be taken into account
when interpreting the result. Can the test be done immediately or is the patient required to prepare in some way, e.g.
fasting? Is the patient taking any medicines that may influence or invalidate the results, e.g. taking antibiotics prior
to faecal antigen testing for Helicobacter pylori? Does the test need to be performed at a certain time,
e.g. measuring testosterone in the morning? Is the test being done at the right stage of illness, e.g. testing for antibodies
after seroconversion has occurred? A test result returned from the laboratory will usually include a reference range or
a threshold value based on guidelines. However, this result “on paper” does not always represent the clinical significance
of a test, which is only apparent once all other factors for that individual patient have been taken into account.
This article is not intended to be a comprehensive guide, but rather an overview of general concepts to guide clinicians
in considering the wide array of factors that can influence laboratory investigations or interpretation of results.
If you are uncertain about how a test may be affected by a specific factor you have identified, discuss this with the
testing laboratory first, or include this detail on your investigation request form.
Biological variation
There are certain variations in laboratory test results that can be expected due to non-modifiable biological factors,
such as age, biological rhythms and physiological changes during pregnancy. These factors may be controlled for, e.g.
by selecting the most appropriate time in the day, month or year for a test, or may be taken into consideration in the
interpretation of results, e.g. different reference ranges or thresholds for clinical significance depending on age, sex
or pregnancy status.
Advancing age
The physiological changes associated with ageing, along with increasing co-morbidities and polypharmacy, mean that older
people are more likely to have test results that fall outside of the normal reference range. For some tests, laboratories
are able to provide an age-adjusted reference range, but for other tests, a result outside of the range in an older patient,
needs to be interpreted in the context of their overall clinical picture.
In many cases, assessing the rate and magnitude of change over time offers more information than interpreting the value
of an individual result. Often the population range of a test shows much more variation than that for an individual patient,
e.g. serum creatinine, liver enzymes. In such cases the patient’s own previous results are a useful baseline.
An example of the effect of age, sex and other variables on interpretation of laboratory results is serum alkaline phosphatase
(ALP), which may be requested as part of liver function tests. The upper reference limit is markedly increased during
puberty as this is the time of maximum bone remodelling. After this period ALP levels fall to a new upper limit through
younger adult life, and then rise again, particularly in females around the time of perimenopause, largely reflecting
an increase in bone turnover at that time. Marked increases in serum ALP may also occur in women in late pregnancy (due
to production of ALP by the placenta), and during other times such as in the weeks after healing of a fracture.
Another example of laboratory values which change with age is lipids. In adults, total cholesterol, LDL and triglyceride
levels increase with age, until approximately age 50 to 60 years in males and age 60 to 70 years in females, when they
begin to decline in most people; triglyceride levels tend to continue to increase in older females.1 Many
laboratories are no longer reporting reference ranges for lipid levels, as this is considered less clinically useful than
treatment targets based on the underlying cause of raised levels and cardiovascular risk.2
When interpreting laboratory results in an older patient, ensure that an age-appropriate
reference range is used if available. Normal age-related changes, e.g. deterioration in renal function,
may explain results outside of the reference range in an older patient, but age alone should not be considered
as the only cause of an abnormal result.
Biological rhythms
Many laboratory parameters vary depending on the time of day, week, month or year when they are sampled. Body temperature,
hormone production (e.g. cortisol, testosterone), platelet and cardiac function and cognitive function follow a circadian
(24 hour) rhythm.1 To allow for this effect, some laboratory tests are recommended at specific times of the
day, e.g. testosterone should be sampled between 7 am – 10 am. This is because peak testosterone levels usually occur
in the early morning; evening levels are often substantially (up to 50%) lower, especially in younger males.2 For
most purposes, serum cortisol* should preferably be sampled in the early morning as there is marked diurnal variation,
with early morning levels at least 50 – 100% higher than levels in the late afternoon.2
Testing vitamin D levels (25-hydroxyvitamin D) is seldom necessary, however, if levels are obtained, they need to be
interpreted in the context of the time of year they are sampled. Seasonal variations occur with vitamin D, with the lowest
levels usually observed at the end of winter (in countries with defined seasons, such as New Zealand). For example, if
a patient has a mild vitamin D deficiency at the end of winter, this is likely to be less clinically significant than
a patient with a mild deficiency at the end of summer.
For further information, see “Age-related testosterone decline in
males” Best Tests (Jun, 2012) and “Vitamin D supplementation:
navigating the debate” BPJ 36 (Jun, 2011).
* Serum cortisol levels have a wide reference range, and therefore this is a relatively
inaccurate measure of cortisol excess or deficiency. The dexamethasone suppression test is used to help
exclude Cushing’s syndrome and the Synacthen stimulation test is used to investigate for primary or secondary hypoadrenalism.2
Menstrual cycle
Females who are menstruating have predictable monthly rhythms of FSH, LH, oestrogen and progesterone. In a general practice
setting, these hormone levels may be requested to investigate conditions such as oligo/amenorrhoea or sub-fertility. The
interpretation and meaningfulness of the results is dependent not only on the value, but on the stage in the cycle when
the hormone was measured.
In a normal menstrual cycle LH levels peak mid-cycle to trigger ovulation. FSH levels also peak mid-cycle. Oestradiol
(the predominant oestrogen in women who are ovulating) is highest prior to ovulation, and then reduces if fertilisation
does not occur. Unless ovulation is being investigated, levels of these hormones are best measured early in the menstrual
cycle.2
N.B. Investigation of oestradiol levels is not useful in women who are taking oestrogen-containing oral contraceptives
as this suppresses the pituitary ovarian axis. LH and FSH are also suppressed in women taking depot progesterone.2
Progesterone levels peak in the second phase of the menstrual cycle, after ovulation has occurred, to prepare the endometrium
for implantation of an embryo. Progesterone levels are sometimes measured to establish if ovulation has occurred; usually
seven days before the expected date of menstruation, i.e. approximately day 21 if the woman has a regular 28 day cycle.
Oestradiol levels decrease and FSH levels increase during menopause, but monitoring these hormone levels is not always
reliable in predicting when a woman is entering menopause, as fluctuations occur with varying ovarian activity.
For further information, see “Reproductive hormones: the right test
at the right time, for the right patient” Best Tests (Feb, 2013).
Pregnancy
Physiological changes during pregnancy result in alterations in many laboratory parameters, such as blood volume, liver
and renal function and hormone levels (Table 1, over page). Reference ranges for different stages of pregnancy are available
for some laboratory tests, however, these ranges are often not as well defined as the general reference range. In addition,
pregnancy-related changes, such as alterations in binding proteins, can affect assays differently, e.g. free hormone levels
can be assay dependent. Therefore caution is recommended in interpreting results based on reference ranges and the laboratory
should be contacted if there is any doubt.
When requesting a laboratory investigation in a woman who is pregnant, note the gestational week on the laboratory
request form.
Table 1: Examples of laboratory values that change with pregnancy1, 2, 3
| Increases |
Decreases |
| Alpha-fetoprotein (AFP); peaks in third trimester |
Haemoglobin; due to haemodilution caused by greater blood volume |
| Alkaline phosphatase (ALP); up to four fold increase in third trimester |
Ferritin; decreases as pregnancy progresses |
| Blood volume (mean plasma volume); increases by 30–50% |
FT4; may decrease slowly in late pregnancy (can be assay dependent) |
| Lipids; up to 40% increase in cholesterol, triglyceride levels can markedly increase in some women (due to the
effect of oestrogen) |
Prothrombin and partial thromboplastin times |
| Creatinine clearance; glomerular filtration rate increases 40-60% (eGFR cannot be reliably calculated) |
TSH; decreases first trimester, then returns to normal (due to the effect of hCG) |
| ESR; increasing to 30–60 mm/h as pregnancy progresses |
Sodium; slight decrease due to changes in blood volume and fluid homeostasis |
| Hormones; oestrogen, testosterone, progesterone, human chorionic gonadotrophin (hCG), prolactin |
|
| Iron binding (transferrin levels); significant increase even in a non-iron deficient woman (due to the effect
of oestrogen) |
|
| White blood count; may increase to 15–18 x109/L |
|
Reference range
A reference range for a laboratory test is a statistically-derived numerical range of results that is obtained by testing
a sample of “healthy” individuals. Defining “healthy”, however, is not straightforward, and depends on a wide range of
factors and assumptions; in many cases to define a range simply using “perfectly healthy” patients would make it unrealistic
and unusable.
The range is also commonly assumed to have a Gaussian distribution, in that 68% of values lie within one standard deviation
(SD) of the mean value, 95% within two SDs and 99.7% of values within three SDs.1 However, many ranges do not have a Gaussian
distribution, but rather the upper end of the distribution is skewed. In these cases the range can be derived either by
log transformation of the data, or simply by identifying the relevant 2.5th and 97.5th percentiles in the population being
studied (with attempts to exclude those patients likely to have an underlying pathology by clinical, laboratory and statistical
means).
Reference ranges for laboratory results usually include two SDs from the mean value meaning that one in 20 “healthy”
individuals will have a test result outside the reference range.1 Reference ranges vary between laboratories, and can
change if new evidence becomes available. The upper and lower limits of the range are not absolute and do not define “normal”
and “abnormal”, but are points at which the probability of clinical significance tends to increase.
Some “reference ranges” are based on recommendations from international bodies for optimising patient outcomes, rather
than on a population statistical distribution. For example, the upper reference limit for TSH in early pregnancy is based
on guidance statements from the Endocrine Society and the American Thyroid Association. The recommended limit for serum
uric acid in patients taking uric acid lowering treatment (0.36 mmol/L) is based on the European League against Rheumatism
(EULAR) guidelines.
Interpretation of a result outside the stated reference range is therefore very dependent on the clinical background
of the patient, the pattern of other abnormalities, and the clinical question(s) being asked.
Individual variations
A patient’s diet, health status and lifestyle factors can all have a pre-analytical influence on laboratory parameters.
Diet and nutritional status
Fasting, calorie restriction, food exclusion diets, malnutrition and dehydration can all affect laboratory results.
The significance of some laboratory tests is dependent on controlling dietary factors, e.g. ensuring there is sufficient
gluten in the diet for at least several weeks prior to serology investigations for coeliac disease, or fasting prior to
assessing the effect of intervention in a patient with previously high triglyceride levels. In other scenarios, assessing
dietary factors can help to interpret unexpected laboratory results, e.g. a vegetarian or vegan diet can result in decreased
levels of vitamin B12, a low carbohydrate diet can cause increased ketone levels (as part of urinalysis) and a high protein
diet can result in increased uric acid levels.
Fasting for 12 hours prior to laboratory testing may be helpful or even necessary, depending on the
clinical scenario, to get the most accurate result for the following tests, which are affected by the ingestion of certain
foods:4
- Glucose; only if indicated – for most patients HbA1c is now recommended as the test of first choice for
the diagnosis and monitoring of type 2 diabetes, and does not require fasting
- Triglycerides; for most patients fasting is not required for lipid testing, but may be useful for monitoring in people
with high triglyceride levels
- Uric acid; fasting is not usually required in practice to get accurate results but the effect of recent dietary intake
may help to explain unexpected results
- Creatinine; a recent meal with high meat content can have a significant influence on serum creatinine, and this should
be considered when monitoring eGFR5
Sustained low caloric intake and starvation can result in numerous changes to laboratory parameters
such as glucose, thyroid function, electrolytes, liver function, renal function and lipids.4 Uric acid levels
may be increased as a result of ketonaemia (causing reduced clearance).3
Malnutrition has varying effects on laboratory results, depending on the nature of the patient’s nutritional
status. Malnutrition is classically thought of as a deficiency of protein and energy, with or without micronutrient deficiencies.
However, malnutrition may be defined as under-nutrition, over-nutrition or deficiency of specific nutrients. Malnutrition
should be considered as a cause for results such as decreased ferritin, folate and vitamin B12 levels.
Dehydration can be considered as a cause of sodium and potassium imbalances, and can also
affect numerous other indices, such as creatinine and urea, albumin, lipids and haematology indices.
For further information see:
“A primary care approach to sodium
and potassium imbalance”, Best Tests (Sep, 2011) and
“Strategies to improve nutrition in elderly people”, BPJ Special Edition (May, 2011).
Caffeine
The effect of caffeine on laboratory parameters has not been fully studied. It has a short half life of three to seven
hours, but this varies among individuals.4 Caffeine intake causes transient increases in blood glucose levels
and impairs glucose tolerance.3, 4 It can also affect other specialised investigations such as interpretation
of metanephrines when investigating hypertension.
Alcohol
The effect of alcohol consumption on laboratory investigations depends on the duration and extent of use. Acute (transient)
effects of alcohol consumption (within two to four hours) include decreased serum glucose and increased plasma lactate
with a reduction in urinary uric acid excretion due to the inhibition of hepatic gluconeogenesis.3
Chronic effects of alcohol consumption on laboratory investigations include:2
- Elevated gamma glutamyl transferase (GGT) and mean cell volume (MCV) which are commonly used to test for excessive
alcohol consumption
- Elevated aspartate aminotransferase (AST), alanine aminotransferase (ALT) and AST/ALT ratio
- Elevated triglyceride levels
- Elevated uric acid and ferritin levels due to fatty liver and alcoholic hepatitis
- Elevated creatine kinase due to alcoholic myopathy
- Other haematological abnormalities, e.g. anaemia and thrombocytopenia
Alcohol consumption can also contribute to vitamin and mineral deficiencies due to replacement of food with alcohol
or as a result of interference of absorption of vitamin and minerals, e.g. decreased folate, vitamin A, vitamin B and
calcium levels.
Timing of investigation in relation to stage of illness
The significance of an investigation can be dependent on when the sample was taken in relation to the stage of the disease
process. The stage of illness can also influence the selection of the most appropriate investigation. For example, a serology
test for syphilis may be falsely negative if the sample is taken too early after exposure, and therefore seroconversion
has not yet occurred. Different types of serology test will provide information about active or past infection.
Acute illness can also affect the result of some investigations, e.g. ferritin is an acute phase protein and levels
can be increased by inflammation and infection, as well as chronic disease and malignancy.
Tobacco smoking
Regular smoking and exposure to nicotine can have both acute and chronic effects on laboratory investigations, although
the mechanisms behind these changes are not fully understood. Within one hour of smoking one to five cigarettes, plasma/serum
concentrations of fatty acids, adrenaline, glycerol, aldosterone and cortisol are increased.3 People who are
chronic smokers may have persistent increases in leukocyte counts, heavy metals, lipoproteins, tumour markers and haematocrit
(PCV), and decreases in the activity of some enzymes (e.g. angiotensin-converting enzyme).3
Exercise
The effect of exercise on laboratory parameters is dependent on the health status of the patient, air temperature during
exercise and intake of food and water during or following exercise.4 Extreme exercise or vigorous exercise
in a person unaccustomed to this level of activity can result in changes to some laboratory parameters. For example, the
most common cause of elevated creatine kinase (CK) levels is exercise. Intense exercise can cause an elevation in CK levels
for several days to a week. Well-muscled people often have CK levels persistently above normal.2
Thyroid function is also known to be altered in people undergoing high-intensity exercise. For example, anaerobic exercise
increases TSH and FT4 levels, but decreases FT3.4 Liver function (AST and to a lesser extent ALT) tests can
increase after exercise. Transient proteinuria and haematuria are also common after exercise, but usually resolve after
a few days. Other analytes that can be increased by exercise include urea, creatinine, lactate dehydrogenase, prothrombin
time, and D-dimer levels. Fibrinogen and the activated partial thromboplastin time (APTT) can be reduced.4 Most
of these effects are likely to be transitory (e.g. persistent for a few hours to a few days after exercise), but this
depends on individual patient factors.
Medicines
The medicines that a patient is taking can significantly affect some laboratory results, therefore this needs to be
taken into consideration when interpreting results. It is good practice to note the relevant medicines that a patient
is taking on the laboratory request form, especially if they may potentially influence results, e.g. antihypertensives
being taken when investigating secondary causes of hypertension or hormone replacement therapy being taken when requesting
endocrine tests.
Medicines can have a direct effect on the sample or laboratory testing process, causing an inaccurate
result. For example, when investigating for H. pylori, a false-negative result of a faecal antigen test may occur
in patients taking a course of antibiotics or proton pump inhibitors (PPIs) as this would decrease the gastric load of H.
pylori.
Medicines may also cause a biological effect to the patient which would account for an altered result.
For example, some antibiotics (e.g. cotrimoxazole and erythromycin), cardiovascular medicines (e.g. amiodarone and propranolol),
NSAIDs (e.g. piroxicam) and gastrointestinal medicines (e.g. omeprazole) may account for a raised INR result in a patient
taking warfarin who normally has a stable INR.2 Long term use of metformin or PPIs is a possible explanation
for a low vitamin B12 level.2 Many medicines have an effect on the balance of sodium and potassium in the body,
e.g. diuretics may cause hypernatraemia (especially loop diuretics), hyponatraemia (especially thiazides), hyperkalaemia
(especially potassium-sparing diuretics) and hypokalaemia (loop and thiazide diuretics).2
For further information, see: “A
primary care approach to sodium and potassium imbalance”, Best Tests (Sep, 2011).
When monitoring the serum concentration or effect of a medicine, the laboratory test needs to be timed
depending on the drug’s metabolism, e.g. a blood sample for testing lithium levels should be collected 10–14 hours after
the last dose and a sample for testing digoxin should be collected at least eight hours after the last dose.2 When
initiating a patient on warfarin, INR levels should be sampled daily in the morning, after an evening dose of warfarin,
to calculate necessary dose adjustments.2
Analytical variation
Analytical variation occurs due to imperfections in testing methods and equipment, which may cause analyte values to
be slightly different each time they are measured. Modern testing methods and laboratory equipment mean that analytical
variation is usually less of a factor in differing test results than biological variation. Ideally the variation in measurement
(expressed as analytical coefficient of variation, or CVa) should be less than half the individual patient biological
variation of the analyte in question (CVi).
For further information on variation with specific analytes, see:
www.westgard.com/biodatabase1.htm
Collection, storage and transport of samples
If a sample is being collected at the practice, it is important to be familiar with the type of collection container
and sample medium that is required by the laboratory for the specific test, as this can affect results, sometimes markedly.
For example, a swab for PCR testing for pertussis should be transported in a dry tube or a tube with universal viral transport
medium, but not in a tube with charcoal transport medium (which is acceptable for swab culture).
Other examples of collection or transport requirements for optimal test results include:
- Blood samples for coagulation studies, including platelet count, D-dimer, prothrombin time, APTT and fibrinogen,
should be transported to laboratory within four hours of collection6
- Urine specimens for culture should be stored in a fridge prior to transportation to reduce the rate of multiplication
of microorganisms
- Samples for glucose analysis should be separated as soon as possible after collection; this applies even with samples
collected in fluoride or oxalate collection tubes, as reduction in glucose concentration still occurs for
60–90 minutes
- Samples for potassium or phosphate should not be left overnight, especially in the fridge, as results can be markedly
altered, e.g. late evening collection with delayed transport to the laboratory
- Faecal samples for culture and microscopy should preferably be transported to the laboratory within four hours
- Semen samples for fertility testing should be kept by the patient at body temperature, e.g. by storing in a clothing
pocket, and transported to the laboratory within one hour of collection.2 The same level of urgency is not
required for post-vasectomy semen analysis.
Refer to your local laboratory for specimen collection requirements. Some providers have online resources, e.g.
www.labtests.co.nz/referrers/tests/collection-guide
Haemolysis
Haemolysis is the destruction of red blood cells, resulting in release of haemoglobin and cellular constituents, e.g.
potassium, into the plasma. It is a cause of inaccurate blood test results and can occur either in vitro during
collection, storage or transportation of a blood sample or in vivo resulting in haemolytic anaemia if severe.
In vivo and in vitro haemolysis have different analytical features, e.g. haptoglobin levels
are normal with in vitro haemolysis, therefore analysis can reveal if haemolysis occurred during the collection
of the blood sample or if it was already present.2 The possibility of haemolysis should usually be noted
on results by the laboratory staff, including the likely degree of interference and reliability of the result.
If the haemolysis is severe, a recollection of the sample may be advised.
Some of the analytes that can be affected when in vitro haemolysis has occurred include:2
- Elevations in potassium, AST (ALT is less affected), lactate dehydrogenase, phosphate
- Reductions in bilirubin, troponin T, insulin
Standardised sample collection and transport processes can help to prevent in vitro haemolysis, including:3
- Allowing alcohol to dry completely when it is used for skin sterilisation prior to venepuncture
- Not leaving a tourniquet on for longer than two minutes
- Using an appropriately sized needle for collection (20 – 22 gauge needles can be used for most routine collections)
- Not removing the needle from the vein if the vacuum tube is attached
- Not exposing the specimen to extremes in temperature
- Avoiding vigorous mixing or shaking of tubes
- Avoiding delay in sending samples to the laboratory