Testing people with conditions suggesting HIV could pick up more recent infections, Europe-wide study finds

A pilot study in 14 European countries, in which groups of patients presenting with one of eight conditions strongly associated with HIV were routinely tested for HIV in various medical settings, found an HIV prevalence rate of 1.8%. This is far in excess of national rates and of the 0.1% prevalence rate suggested by the US Centers for Disease Control for routine HIV testing of the population, and suggests targeting people with these so-called indicator conditions (ICs) for HIV testing could be an efficient way of detecting infection.

The study found relatively high CD4 counts and a high prevalence in patients presenting with mononucleosis-like illness. These symptoms are usually caused by Epstein-Barr virus but may also be due to HIV seroconversion. This suggests that testing for indicator conditions might be a good way of detecting more people in early HIV infection.

The study

The HIDES I study arose from the HIV in Europe conference in 2007, which debated how to improve HIV testing rates and reduce the amount of undiagnosed HIV in the region.

One method suggested was to draw up a list of indicator conditions (ICs) associated with HIV probability of HIV infection and recommend that people presenting of these should be routinely tested.

HIDES I drew up a shortlist of eight ICs and sent out a call to healthcare centres in Europe to participate in the study to test the feasibility and acceptability of routine IC-driven HIV testing.. This would involve them routinely test all patients consecutively presenting with a specific IC.

The eight conditions chosen were:

• Acute STIs
• Hepatitis B or C
• Malignant lymphoma of any type
• Anal or cervical intraepithelial neoplasia (HPV disease) grade 2 or above
• Unexplained low platelets (thrombocytopenia) or neutrophils (neutropenia) for more than four weeks
• Herpes zoster (shingles) in people under 65
• Seborrhoeic dermatitis or exanthema (sudden rash)
• Mononucleosis-like illness.

Results: demographics

Seventeen centres in 14 European countries spread over the whole continent from UK to Ukraine participated in HIDES I. They conducted 39 separate surveys on HIV testing in patients presenting with ICs between September 2009 and February 2011. There were one to six surveys conducted in each centre and between three to six surveys comducted for each indicator condition.

Altogether, 3588 patients were tested for HIV. They had an average age of 36, which varied according to condition from 24 for mononucleosis to 53 for lymphoma. Fifty-five per cent were men and a quite high proportion (36%) had had a previous HIV test.

Sex, age, sexuality and previous testing rates varied by European region. Patients were youngest (average 33) in northern Europe (UK, Netherlands, Denmark, Sweden) and oldest (43) in southern Europe (Italy and Spain). Roughly two-thirds were male in all European regions apart from northern Europe, where 44% were male.

Only two per cent of patients in eastern Europe (Belarus, Ukraine, Bosnia, Croatia, Poland) said they were gay compared with approximately 40% in south and west central Europe (Germany, Austria, Belgium) and 45% in northern Europe; and only 13% in eastern Europe had had a previous HIV test compared with 36% in southern, 45% in west central, and 60% in northern Europe.

Results: HIV diagnoses

There were 66 new HIV diagnoses (1.8% of patients). Eighty-three per cent of those diagnosed were male, 58% of them were gay, 42% heterosexual and 9% had injected drugs. The prevalence of newly-diagnosed HIV varied widely from 0.6% in northern Europe to 6.5% in southern.

It also varied widely per IC. Not unexpectedly the highest HIV diagnosis rate was in surveys involving STIs (4%) but it was not much lower in mononucleosis-like illness (3.8%), neutro/thrombocytopenia (3.2%), shingles (2.9%) and dermatitis/rash (2.1%).

In contrast HIV prevalence was only 0.4% in cervical/anal dysplasia and hepatitis B/C and 0.3% in lymphoma, possibly because the individuals with these diseases who have HIV are more likely to have already been diagnosed. These rates are still above the 0.1% rate considered cost-effective for routine testing, though.

Over the whole group, more than half of those diagnosed with HIV(52%) had been tested for it before, with a median time from their previous HIV test of 19 months. Their average CD4 count was 400. This may indicate that offering an HIV test as standard when ICs are diagnosed may detect more cases of early HIV infection, though presenter Ann Sullivan of London's Chelsea and Westminster Hospital did not split this analysis into European regions. It is possible that the relatively high amount of recent infections may be driven by tests given to people presenting with STIs who might have tested anyway.

HIDES 1 found that 20% of patients had presented to healthcare with one of the ICs in the preceding five years and 4.6% had presented more than once.

Next steps

The HIDES project now intends to conduct a larger HIDES 2 study involving 12 indicator conditions, in order to establish better HIV prevalence figures for each. It also intends to conduct a separate audit of whether patients in Europe who present with ICs are offered HIV tests. Finally they intend to develop, in collaboration with the European Centre for Disease Control and the World Health Organization, a set of guidelines for HIV testing targeted at indicator conditions.

HIDES investigator Jens Lundgren commented from the audience that although HIDES 1 was a pilot study, its results were important because many physicians had been sceptical that using ICs to guide HIV testing would find higher than average rates of HIV or be cost-effective. These results showed that it was an efficient way to find undiagnosed HIV.