We need to remember that there is always a patient at the receiving end of a test result, and that, wherever possible, the result should have a positive bearing on the management of the patient.

However some results can cause potential harm to the patient. How can this be?

A classic example of this is the urinary antigen test for Legionella pneumophilia. At first glance, this test, easily obtained and giving a result within 20 minutes, seems to do the job just fine.

But all is not quite as it seems….

The urinary antigen test for L pneumophilia only picks up serogroup 1. What about all those other serogroups of L. pneumophilia? It also has a sensitivity of about 70-75%. So 25-30% of true L pneumophilias (serogroup 1) are being missed. And finally, in NZ at least, the vast majority of Legionella disease is caused by Legionella longbeachae, for which the urinary antigen test is no good whatsoever.

The big worry is that a negative urinary antigen test in a sick patient will lead to antibiotic cover for Legionella being discontinued, and I have seen this happen quite a few times over the years, even in the ICU setting. We see this happen because the recipient of the result often takes it for gospel, without considering the whole picture of both the result and the patient, the sensitivity, specificity, PPV, clinical picture etc.

With the increasing availability of PCR testing for Legionella, maybe the days of urinary antigen tests for Legionella pneumophilia serogroup 1 are numbered, in much the same way as rapid turnaround PCR for RSV and Influenza are gradually pushing out their rapid antigen equivalents for these infections.

It is always a concern when a laboratory test can potentially do more harm than good, and I think this sometimes happens with urinary antigen for Legionella.

We have to be brave enough to both embrace and introduce new tests, and at the same time get rid of the old…

Michael