The Western Blot test is a kind of ELISA test that distinguishes between antibodies that recognise different antigens in the test system.

So instead of serum being put (say) into a hollow well on a plate, it is spread over a nitrocellulose-coated plate to which the antigens have been bound in different places.

Because each plate is slightly different in the position of these bands of antigen, the test must use control sera and the results are read by looking at the depth of the colour change on each band.

The plates are expensive to make, and the test can give indeterminate results, when just one or two bands change colour or the colour change is not very marked. However, if the result shows that multiple HIV antigens are recognised by antibodies in the test sample, then the test is reasonably conclusive.

It is also possible to 'print' different antigens onto a strip in different places, to create a test that looks like a Western Blot and is used in the same way, but could distinguish between different antigens that wouldn't be separated on a real Western Blot - for example, because they were of the same molecular weight.

Limitations of Western Blot

Western Blot relies on a subjective judgement of the intensity of the bands of antibody which form in response to particular antigens. It also relies upon well–trained technical staff, so false positive results are possible with this method just as with any other laboratory test. Early literature which describes Western Blot as a gold standard for antibody testing should be read with caution.

False positive results could arise as a result of similarities in the molecular weight of proteins derived from cellular material rather than from HIV.

Rare antibody cross–reactions have been documented in individuals with auto–immune illnesses such as lupus and multiple sclerosis, and in the case of people who have recently been vaccinated against hepatitis B, influenza and tetanus. Transient positive HIV antibody reactions after vaccination have been reported, but a positive antibody test for HIV has not been detected upon subsequent tests. Transient false positive results have also been seen in blood dialysis patients and recipients of blood transfusions. It should be emphasised that all these false positive results are isolated reports; in the case of multiple sclerosis, for instance, such reports are vastly outnumbered by studies which have failed to find HIV antibodies in large numbers of MS patients who lack known risks.

Particular antibodies considered to be specific for HIV sometimes appear in uninfected people as a consequence of other infections or auto–immune problems. For instance, 13% of people with warts were shown in one study to have antibodies to p24 but not to any other HIV antigens. In the same study 41% of a group of patients with multiple sclerosis were shown to have antibody to p24 but no other antibody (Ranki).

HIV may share similar genetic sequences to human retroviruses which have become encoded into human genes (endogenous retroviruses). These sequences are confined to the gag region alone, and do not include the env and pol regions (Haist). This may be another explanation for the greater frequency of false p24–positive reactions on the Western Blot test.

The p24 reaction may also be partly attributable to cross–reaction with cytomegalovirus (Landini). CMV does not generate cross–reactions to other HIV proteins.

Cross–reactions to other proteins such as gp41 have also been observed, but cross–reactions to all these proteins at once has never been reported.

Thus it is very important that HIV infection only be diagnosed upon the presence of a range of HIV antibodies: antibodies to p24 and p31 (HIV core proteins) are not sufficient without antibodies to either gp41, gp120 or gp160 (envelope proteins).

Western Blot should NOT be used as a stand-alone HIV antibody test, but only as part of a multi-test strategy.