The interpretation of results

It can take considerable effort, and expense, to produce what may seem to be just numbers on pieces of paper or on a computer screen. Understanding what these numbers mean is of crucial importance if the correct diagnosis is to be made, or if the patient’s treatment is to be changed.

How biochemical results are expressed

Most biochemical analyses are quantita- tive, although simple qualitative or semi- quantitative tests, such as those for the presence of glucose in urine, are com- monly encountered methods used for point of care testing. Many tests measure the amount of the analyte in a small volume of blood, plasma, serum, urine or some other fluid or tissue. Results are reported as concentrations, usually in terms of the number of moles in one litre (mol/ L) (Table 3.1).

The concept of concentration is illus- trated in Figure 3.1. The concentration of any analyte in a body compartment is a ratio: the amount of the substance

Fig 3.1 Understanding concentrations. Concentration is always dependent on two factors: the amount of solute and the amount of solvent. The concentration of the sugar solution in the beaker can be increased from 1 spoon/ beaker (a) to 2 spoons/beaker by either decreasing the volume of solvent (b) or increasing the amount of solute (c)

dissolved in a known volume. Changes in concentration can occur for two reasons: 

- The amount of the analyte can increase or decrease. 

- The volume of fluid in which the analyte is dissolved can similarly change. 

Enzymes are not usually expressed in moles but as enzyme activity in ‘units’. Enzyme assays are carried out in such a way that the activity measured is directly proportional to the amount of enzyme present. Some hormone measurements are expressed as ‘units’ by comparison to standard reference preparations of known biological potency. Large mole- cules such as proteins are reported in mass units (grams or milligrams) per litre. Blood gas results (PCO2 or PO2) are expressed in kilopascals (kPa), the unit in which partial pressures are measured.

Variation in results

Biochemical measurements vary for two reasons. These are described as ‘analyti- cal variation’ and ‘biological variation’. Analytical variation is a function of ana- lytical performance; biological variation is related to the actual changes that take place in patients’ body fluids over a period of time

Laboratory analytical performance

A number of terms describe biochemi- cal results. These include:

- precision and accuracy

- sensitivity and specificity

- quality assurance

- reference intervals.

Fig 3.2 Precision and accuracy.

Precision and accuracy 

Precision is the reproducibility of an analytical method. Accuracy defines how close the measured value is to the actual value. A good analogy is that of the shooting target. Figure 3.2 shows the scatter of results which might be obtained by someone with little skill, compared with that of someone with good precision where the results are closely grouped together. Even when the results are all close, they may not hit the centre of the target. Accuracy is there- fore poor, as if the ‘sights’ are off. It is the objective in every biochemical method to provide good precision and accuracy. Automation of analyses has improved precision in most cases. 

Analytical sensitivity and specificity 

The analytical sensitivity of an assay is a measure of how little of the analyte the method can detect. Analytical specificity of an assay relates to how good the assay is at discriminating between the requested analyte and potentially inter- fering substances. These terms describ- ing the analytical properties of tests should not be confused with ‘test’ spe- cificity and sensitivity, as applied to the usefulness of various analyses (see below).

Quality assurance

Laboratory staff monitor performance of assays using quality control samples to give reassurance that the method is per- forming satisfactorily with the patients’ specimens. Internal quality control samples are analysed regularly. The expected values are known and the actual results obtained are compared with pre- vious values to monitor performance. In external quality assurance programmes, identical samples are distributed to labo- ratories; results are then compared

Fig 3.3 (a) Overlap of biochemical results

in health and disease. (b) and (c) The effect of changing  the diagnostic cut-off on test specificity and sensitivity.

Reference intervals Analytical variation is generally less than that from biological variation. Biochemi- cal test results are usually compared to a reference interval chosen arbitrarily to include 95% of the values found in healthy volunteers (Fig 3.3). This means that, by definition, 5% of any population will have a result outside the reference interval. In practice there are no rigid limits demarcating the diseased popula- tion from the healthy; however, the further a result is from the limits of the reference interval, the more likely it is to indicate pathology. In some situations it is useful to define ‘action limits’, at which appropriate intervention should be made in response to a biochemical result. An example of this is plasma cholesterol.

There is often a degree of overlap between the disease state and the ‘normal value’ (Fig 3.3). An abnormal result in a patient who is subsequently found not to have the disease is called a ‘false positive’. A ‘normal result’ in a patient who has the disease is a ‘false negative’.

Specificity and sensitivity of tests 

The specificity of a test measures how commonly negative results occur in people who do not have a disease. Sensitivity is a measure of the incidence of positive results in patients who are known to have a condition. As noted above, the use of the terms specificity and sensitivity in this context should not be confused with the same terms used to describe analytical performance. An ideal diagnostic test would be 100% sen- sitive, showing positive results in all dis- eased subjects, and 100% specific, with negative results in all persons free of the disease. Figure 3.3 shows the effect of changing the ‘diagnostic cut-off value’ on test specificity and sensitivity. 

Biological factors affecting the interpretation of results 

The discrimination between normal and abnormal results is affected by various physiological factors that must be con- sidered when interpreting any given result. These include: 

- Sex. Reference intervals for some analytes such as serum creatinine are different for men and women.

- Age. There may be different reference intervals for neonates, children, adults and the elderly.

- Diet. The sample may be inappropriate if taken when the patient is fasting or after a meal.

- Timing. There may be variations during the day and night.

-  Stress and anxiety. These may affect the analyte of interest.

- Posture of the patient. Redistribution of fluid may affect the result.

- Effects of exercise. Strenuous exercise can release enzymes from tissues.

- Medical history. Infection and/or tissue injury can affect biochemical values independently of the disease process being investigated.

- Pregnancy. This alters some reference intervals.

- Menstrual cycle. Hormone measurements will vary throughout the menstrual cycle.

- Drug history. Drugs may have specific effects on the plasma concentration of some analytes

Other factors

When the numbers have been generated, they still have to be interpreted in the light of a host of variables. The clinician can refer to the patient or to the clinical notes, whereas the biochemist has only the information on the request form to consult. 

The clinician may well ask the follow- ing questions on receiving a biochemistry report:

- ‘Does the result fit with the history and clinical examination of the patient?’

- ‘If the result is not what I expected, can I explain the discrepancy?’

- ‘How can the result change my diagnosis or the way I am managing the patient?’

- ‘What should I do next?’

What is done in response to a biochemistry report rests with the clinical 

judgement of the doctor. There is a maxim that doctors should always ‘treat the patient, rather than the laboratory report’. The rest of this book deals with the biochemical investigation of patients and the interpretation of the results obtained.

Clinical note

It is important to realize that an abnormal result does not always indicate that a disease is present, nor a normal result that it is not. Beware of over-reacting to the slightly abnor- mal result in the otherwise healthy individual.