BHIVA: One in 40 people presenting to A&E departments with ‘flu’ symptoms has HIV

An anonymous survey in Brighton of male patients aged 18-50 reporting to GPs and to hospital accident and emergency (A&E) departments with symptoms suggestive of HIV infection has found an overall HIV prevalence of one per cent in those previously undiagnosed with HIV. However it found a higher prevalence of 2.5% amongst the patients turning up at A&E, suggesting that, at least in a high-prevalence town like Brighton, there is scope for ‘opt-out’ HIV testing to be introduced in emergency settings.

Dr Kate Nambiar from Brighton and Sussex University Hospital told the conference that a study tested residual blood samples left over from patients who had presented to local GP practices and emergency departments between January 2006 and June 2007 and who had been tested for Full Blood Count or for glandular fever.

This was an anonymous unlinked survey: although details such as the sex and (known) HIV status of the patient were kept, samples were anonymised and ones that tested positive could not be traced back to individuals.

In order for samples to be tested, patients had to have presented with flu-like symptoms, muscle and joint aches, fever, night sweats, rash or mouth ulcers or a combination of other symptoms suggestive of primary HIV infection.

Samples were tested for HIV antibodies and the HIV p24 protein (which appears earlier than HIV antibodies in infected people). Pooled samples were also tested for HIV RNA (i.e. given a viral load test) in order to identify people in primary infection before seroconversion.

Samples positive for HIV were then sent on to the Health Protection Agency’s laboratory for confirmation with a Western Blot HIV test and to be tested with the STARHS assay, which can detect infections less than six months to a year old.

A total of 686 samples were tested and of these 7% tested HIV positive. Eliminating patients known to have HIV, this left 1% of patients (seven samples) who tested positive for HIV and were undiagnosed. One of these samples tested positive for primary HIV infection. It was recorded that 33 of the 686 patients had taken a same-day HIV test, but all were negative.

Seventy-one per cent of samples were collected from GPs and 29% from A&E departments; however only two of the positive samples were from GPs and the other five were from A&E. This meant that the proportion of undiagnosed patients attending GPs was only 0.4% (one in 250) but the proportion attending A&Es was 2.4% (one in 40).

Dr Nambiar commented that the study had found HIV in a considerably lower than expected proportion of patients presenting with symptoms suggestive of primary HIV infection (PHI) than had some previous US surveys; these had found PHI levels in undiagnosed symptomatic patients of up to 0.8%.

Nambiar commented that the suspected reason was that patients presenting with flu-like symptoms, particularly at GPs, were not undergoing any blood tests at all: “They may just be told ‘it’s only a virus’, and sent home with a prescription for paracetamol,” she said.

The study therefore did not support the use of pooled RNA testing as a means of picking up on primary HIV infection in the community.

However, she added, given that HIV screening is probably cost-effective if done in populations where the background undiagnosed HIV prevalence is 0.05% (one in 2,000) or more, the relatively high prevalence of undiagnosed infection in symptomatic individuals presenting to A&E suggested that, even if their symptoms were not caused by HIV, in Brighton at least, “there is a case for routine HIV testing in accident and emergency”.