HIV incidence in gay men unchanged in England and Wales, despite more testing

Treatment on diagnosis may be needed, say researchers
This article is more than 11 years old. Click here for more recent articles on this topic

A paper in The Lancet Infectious Diseases by scientists from the UK’s Medical Research Council and the Health Protection Agency (HPA) has calculated that the number of gay men in England and Wales who become infected with HIV each year remained unchanged between 2001 and 2010. This is despite a considerable increase in testing and, they estimate, a 40% reduction in the proportion of gay men with HIV who are undiagnosed.

The paper concludes that, in England and Wales at least, the proportion of gay men with HIV who are on treatment and with undetectable viral loads is currently too low to bring about a decline in annual HIV incidence in this population. This is in contrast to declines in diagnosis, and claims of declines in incidence, seen in places such as San Francisco, the province of British Columbia in Canada, and some locales in South Africa.    

As well as extending HIV testing to non-traditional settings and urging gay men to test more frequently, the authors conclude that “the initiation of treatment on diagnosis, regardless of CD4 count might well be necessary to achieve control of HIV transmission”, and welcome the new BHIVA treatment guidelines' recommendation "that clinicians discuss the benefits of early treatment uptake as a prophylaxis to protect sexual partners” as a step towards this.

Calculating incidence

The paper is a mathematical model. It uses available data on diagnoses, CD4 counts at diagnosis, and the proportion of people on antiretroviral therapy (ART) to make estimates of the true annual number of infections (annual incidence) in gay men, the number undiagnosed, average time gap between infection and diagnosis, the distribution of CD4 counts among diagnosed and undiagnosed men and the proportion who are on treatment and with an undetectable viral load.

Glossary

mathematical models

A range of complex mathematical techniques which aim to simulate a sequence of likely future events, in order to estimate the impact of a health intervention or the spread of an infection.

detectable viral load

When viral load is detectable, this indicates that HIV is replicating in the body. If the person is taking HIV treatment but their viral load is detectable, the treatment is not working properly. There may still be a risk of HIV transmission to sexual partners.

UNAIDS

The Joint United Nations Programme on HIV/AIDS (UNAIDS) brings together the resources of ten United Nations organisations in response to HIV and AIDS.

Although mathematical models are always estimates, in this case surveillance data from the UK are of good enough quality to make them quite robust, though because by definition fewer very recent infections are diagnosed, incidence estimates for the last two years are less certain than for previous years.

The incidence rate is not the same as the new diagnosis rate in HIV, because of the time lag between infection and diagnosis. If the number or frequency of HIV tests go up, the number of diagnoses will tend to go up, since more long-term undiagnosed infections will be identified. The researchers got round this problem by using CD4 count at diagnosis – available for the majority of diagnosed people in England and Wales – as a surrogate for the time delay between infection and diagnosis, given that CD4 counts in people with untreated HIV tend to decline at an even rate over time.   

Results – diagnosed and undiagnosed

The number of diagnoses in gay men in England and Wales increased from about 1800 in 2001 to 2600 in 2010. However by adjusting this for CD4 count at diagnosis, the researchers estimated that the true annual total of HIV infections in gay men had remained virtually unchanged, from 2200 in 2001 to about 2300 in 2010. There was an increase in incidence to about 2700 a year in 2003-4, due to increased rates of sex without condoms in gay men, but this has reduced since.

This reduction is due, the researchers say, to more gay men taking tests and to a shorter period between HIV infection and diagnosis. The number of HIV tests taken by gay men in sexual health clinics has grown nearly fourfold, from 16,000 in 2001 to 59,300 in 2010. As a result, the estimated time between infection and diagnosis has shrunk from four years to 3.2 years during this time, and the proportion of gay men with HIV who are undiagnosed from 37 to 22%.

The reason it has not shrunk more, say the authors, is due to gay men not testing often enough. Last year, study co-author Valerie Delpech of the HPA told the IAPAC Prevention Summit that only an estimated 10 to 15% of gay men took an HIV test every year, and that two-thirds of gay men who had had a test at a clinic had, two years later, not returned to that clinic for another one.     

Because there are (as of 2010) 3.2 years’ worth of undiagnosed infections in the population, the total number of gay men with HIV who are undiagnosed in England and Wales was estimated as 7690 in 2010. This was only a small increase from 7370 in 2001 and represents a 16% decline from 9140 in 2004-5, again due to more testing.

The proportion of gay men with HIV who are undiagnosed has gone down by 40% while the number has scarcely changed because total HIV prevalence and the number of UK gay men living with HIV has grown over the same period.

Results – implications for treatment

In 2001, at HIV diagnosis, about 65% of gay men had a CD4 count under 500 cells/mm3, 40% under 350 cells/mm3, and 18% under 200 cells/mm3. Ten years later, the proportion in these three categories had only fallen by about 5%. This means that less than 40% of gay men would currently be advised, under treatment guidelines, to begin taking antiretroviral therapy (ART) for treatment reasons as soon as they are diagnosed.

The researchers calculated that, because more undiagnosed infections are recent ones, only 20% of undiagnosed gay men had a CD4 count under 350 cells/mm3 and only 45% under 500 cells/mm3. Further decreasing the proportion of gay men with HIV who are undiagnosed, and raising or abolishing the CD4 threshold for treatment initiation, would therefore have considerable cost implications for the National Health Service in England and Wales. 

Conclusions

In many ways, the UK’s response to HIV has been excellent. The proportion of gay men with a CD4 count under 350 cells/mm3 who are on ART has increased from 75% in 2001 to 84% in 2010; 65% of all patients in care, including the untreated, have undetectable viral loads; and annual loss to follow-up of those attending care is under 5%.

In the US, in contrast, it is estimated that there are more gay men who are diagnosed but not taking ART than there are undiagnosed, and that only 28% of people with HIV are virally suppressed. But gay men in other countries test more frequently: as an accompanying editorial by Reuben Granich of UNAIDS points out, the 22% of gay men who remain undiagnosed in the UK is not as good as an estimated 14% in Vancouver and only 6% in San Francisco.

Because most of those with detectable viral loads in the UK are undiagnosed, it is estimated by the HPA that up to 50% of HIV infections in gay men here could be being transmitted by men in primary HIV infection and another 35% by undiagnosed men with long-term infection. The authors conclude that treatment initiation at diagnosis, earlier, more targeted testing, and better primary HIV prevention all need to be part of any national HIV prevention plan for England and Wales.

References

Birrell PJ et al. HIV incidence in men who have sex with men in England and Wales 2001-2010: a nationwide population study. The Lancet Infectious Diseases, early online edition: http://dx.doi.org/10.1016/S1473-3099(12)70341-9. See abstract here. 2013.

Granich R HIV in MSM in England and Wales: back to the drawing board? The Lancet, early online edition: http://dx.doi.org/10.1016/S1473-3099(13)70035-5. See first few lines here. 2013.