How long have I got, doc?

Can people with HIV really live as long as anyone else? Gus Cairns investigates

Last month we featured a news report about a couple of studies,^1,2 presented at a recent conference, which found that certain groups of HIV-positive people could expect to live as long as comparable HIV-negative people.

A study from the Netherlands looked at people who were diagnosed between 1998 and 2007 and then excluded those who had had to go on treatment within six months of diagnosis. Once this group, who’d mostly have been diagnosed with a low CD4 count, was excluded, then the average remaining life expectancy of someone who was diagnosed at the age of 25 was calculated to be 52.7 years – in other words they would die, on average, at the age of nearly 78: just five months short of a 25-year-old member of the general Dutch population.

This group had a high average CD4 count of 480 cells/mm³, and the study extrapolated life expectancy from an average of just 3.3 years of mortality data, and a maximum of ten; we need to be careful about interpretation because, as we see below, things may take a turn for the worse later.

The other study was an analysis of the COHERE cohort, a group of over 80,000 HIV-positive people from more than 30 European countries. It included all people in COHERE who had started treatment later than 1998, thereby excluding people who had taken pre-HAART drug regimens, who are the ones most likely to have significant drug resistance. It didn’t estimate a life expectancy: instead it calculated something called the standardised mortality ratio (SMR). This is the amount the death rate in a group differs from the death rate in the general population.

The headline finding was a correction to any assumption that most people with HIV are now living normal lifespans. Over the whole group, which included people of every CD4 count, the death rate was 4.4 times what you’d expect to see in 80,000 people of the same age and sex picked at random from the general population.

However, the SMR in men whose current CD4 count was over 500, and who had maintained it for over three years (or just one year if you excluded injecting drug users) was 1.1: statistically, the same as people without HIV.

There were fewer actual deaths in women than men, but mortality in HIV-positive women was twice as high as in women in the general population because the death rate amongst HIV-negative women of similar age is considerably lower than in men.

There’s no doubt that life expectancy for people with HIV, at least in the developed world, has improved vastly since treatment became available. But is it continuing to improve? Will more of us achieve a normal lifespan as time goes on? And what do we need to do to ensure this happens?

Several other studies have addressed this question in the last decade.

One of the problems besetting life-expectancy studies is that very different groups of patients are selected for study. For instance, a study from the USA³ found that, from the point of diagnosis, people with HIV on average lived 21 years fewer than HIV-negative people of the same sex and age. But one-third of this group had a CD4 count under 200 cells/mm³, and many were not accessing health care.

Another problem is that very different measures of mortality are used so it’s not easy to make comparisons. These include absolute mortality, the excess mortality compared with the general population (as in COHERE): the expected average number of years of life lost, given this excess mortality (for instance, smokers will lose an estimated ten years of life to their habit, compared to non-smokers), and life expectancy.

Jonathan Sterne of the Antiretroviral Therapy Cohort Collaboration, author of one of the studies we quote in this piece, says: “Life expectancy is a strange concept in that it extrapolates into the future a present state of affairs. It says: ‘Given the current mortality rates, if nothing changes, how long can people expect to live’?

“It’s rarely going to reflect what people’s average lifespan actually ends up being, because it can’t take account of future developments.

“For instance, the life expectancy of people with HIV may improve, because treatments get better. But on the other hand, it may also unexpectedly decrease if we see in future a sudden increase in the death rate at a certain age, or after a certain time spent taking HIV drugs.”

There’s no doubt that life expectancy for people with HIV, at least in the developed world, has improved vastly since treatment became available. But is it continuing to improve?

One study⁴ of another cohort, called CASCADE, took account of this by observing the excess mortality, compared with the general population, over specific two-year slots post-diagnosis.

This study found that there were 1239 deaths in 7034 patients diagnosed between 1996 and 2006, where in the general population you’d only expect 178.7 deaths. That means over the whole study period deaths in people with HIV were sevenfold higher than they were in HIV-negative people.

However, this excess mortality went down in every period of the study. In 1996-97 people with HIV had 17 times the death rate of the general population. By 2004-2006 it was 3.4 times the rate.

Furthermore, because in this study the date people were infected was approximately known, changes in the number of excess deaths over time could be looked at. 

By the year 2001, people who had been under 35 at diagnosis were no more likely to die than the general population in the first five years after being diagnosed, and by the year 2006 this had extended to people diagnosed under 45. By this time the ten-year death rate was also starting to approach normal among the under-45s.

However, the death rate amongst people diagnosed for 15 years was still very much higher than in the general public amongst all groups: seven times higher in people diagnosed up to the age of 25 (meaning they’d be up to 40 years old now): 5.5 times higher in people diagnosed before 35 (so now in their 50s): and 2.4 times higher in people diagnosed up to the age of 45 and now in their 60s.

Here’s the possible reason why the death rates seen in this study, and the life expectancy one might derive from them, are higher in this study than in the Dutch one that predicted normal life expectancies. If you only look at death rates in people with HIV for the first decade after they test positive, you may miss most of the excess deaths, whereas CASCADE followed some people up to 24 years after infection.

In the CASCADE study the death rate in people with HIV, as measured in 2006, was very little higher than in the general population until about eight to nine years after infection. After this it began to outpace the expected death rate.

Was this because eight to nine years after seroconversion is when HIV starts to make people ill? Or were people with HIV diagnosed before 1998 more likely to have taken suboptimal drug regimens which led to the development of drug resistance?

Kholoud Porter of the UK Medical Research Council says that using ‘time since infection’ rather than age to distinguish different mortality rates was vital because “a 45-year-old may have only just been infected, or they may have had 25 years of living with HIV and HIV medications. If you don’t take time since infection into account you may over- or underestimate mortality.”

Just two weeks after the CASCADE study came out, an even larger study called the Antiretroviral Therapy Cohort Collaboration study⁵ was published, which did extrapolate life expectancy in 43,355 patients from North America and Europe.

It found that deaths had declined, from one death per 60 patients a year in 1996-99 to one per 100 in 2003-05. And it found that life expectancy had increased, from 36.1 years for a 20-year-old in 1996-98 (so they could, on average, expect to live till they were 56) to 43.1 in 2003-5 (living till 63). Thirty-five-year-olds could expect to live till 60 in 1996-99 and 72.3 years in 2003-5. But SMRs were still in the region of six to eight times that of the general population, depending on year of diagnosis.

Conclusions

The lessons to take from all this? Australian HIV Expert Professor David Cooper commented on this study in an editorial.⁶

“These figures will help clinicians raise the hopes and expectations of patients during discussions of life choices and goals,” he said.  “But life expectancy is still not normal: about ten years is shaved off a normal lifespan.”

Life expectancy is continuing to improve in people with HIV, and in certain groups of people who test and access care and treatment promptly it may be approaching that of the general population.

But there is still a long way to go. Those diagnosed late still face a 20-year life deficit.

Life expectancy also doesn’t tell us about quality of life as we age. The next generation of research needs to concentrate on health as well as death statistics: not just whether we will have an old age, but whether we will have a healthy one.