The study looked at mortality rates in 24,768 HIV-positive people
within the Kaiser Permanente system and compared them with ten times that
number of HIV-negative people (257,600) also in the system between the years 1996 and 2011. The
two groups were matched for age (31 at entry to the study) and gender (91%
male). Deaths were ascertained from death certificates and social security
records so could be traced even if people left the Kaiser system.
They were matched approximately for ethnicity: 56% of
HIV-positive and 44% of HIV-negative people were white, 21% versus 25% were
black, and 18% versus 10% were Hispanic. Ethnicity was less well-recorded in
Forty-five per cent of the HIV-positive people had ever
smoked versus 31% of the HIV-negative; 21% of positive versus 9% of negative
had ever had drugs or alcohol problems; and 12% versus 2% had ever had
hepatitis B or C.
In the HIV-positive people, 75% were men who acquired HIV
through sex with other men; 16% were heterosexual men and women; 7% got HIV
through injecting drug use; and 2% through other routes such as occupational
exposure. Forty-six per cent were already on ART when
they joined the Kaiser cohort, while another 40% started during their period in
the study. One in three (35%) did not start ART until their CD4 count was below
200 cells/mm3, while 18% started at CD4 counts over 500 cells/mm3.
The cohort study starts at 1996, which was just before ART
became generally available, so there was a rapid decrease in mortality of HIV-positive people in the first two years of the study; it then continued to decline at a
slower rate from 1998 onwards. In 1996-97 the death rate in HIV-positive people
was 7.08% a year. By 2011 this had declined to 1.05% a year. The equivalent
rates in HIV negative people were 0.44% a year in 1996-97 and 0.38% a year in
What did this do to life expectancy? In 1996-97 the life
expectancy at age 20 of an HIV-positive person was 19 years, in other words they
could only expect to live, on average, in the absence of any improvement in
treatment, till they were 39. By 2011, this had improved to 53 years, i.e. death
on average at 73.
For HIV-negative people, life expectancy at age 20 in 1996-97 was 63 years; by 2011 this had improved to 65 years, i.e. death
on average at 85.
In HIV-positive women, life expectancy improved slightly
less than it did on men. For subgroups, instead of contrasting 1996-97 with
2011, the researchers contrasted life expectancy during the whole period
between 1996 and 2007 with life expectancy from 2008-2011.
For HIV-positive men, life expectancy at age 20 was 37 more
years in the 1996-2007 period and 51 years in 2008-2011; for women it was 38
years between 1996 and 2007 and 49 years in 2008-2011.
The increase in life expectancy in white people was the same
as it was in men. In black people it was lower during both periods and did not
improve as much as in white people (38 years at age 20 in 1997 to 2007 and 46
in 1998-2001 – this was pretty much the same increase as in people who inject drugs); while Hispanic people did rather better with an improvement from 39 to 52
years. In gay men it improved from 40 to 51 years.
One interesting aspect of this study is that death rates and
life expectancy in HIV-positive people has tended to be compared with the
general population’s figures. But of course people with HIV form part of the
general population. Thus, taken over the whole study period, life expectancy in
HIV-positive people at age 20 was 49 years; in HIV-negative people it was 62
years; and in the US general population it was 60 years. This means that if the
Kaiser Permanente HIV-negative population resembles the HIV-negative US general
population, then HIV reduces life expectancy in the US general population by two years.
In subgroups, the difference was bigger: in men the gap
between HIV-negative and general-population life expectancy was five years, in
black people three years, and in Hispanic people six years. However, Kaiser’s
users are not likely to resemble the
general population, so the actual reduction in life expectancy due to HIV in the
general population is likely to be lower than this.
There still remains a gap of 13.1 years between HIV-positive
and HIV-negative life expectancy in this study, and this did not improve
between 2008 and 2011. The researchers then looked at factors that might narrow
this gap. In people who started ART at CD4 counts over 500 cells/mm3,
the life expectancy gap between them and HIV-negative people was 7.9 years,
i.e. it added 5.2 years to an HIV-positive person’s life expectancy. Not having
had hepatitis B or C added 5.9 years; not having had problems with drugs or
alcohol added 6.5 years; and not having ever smoked added 7.7 years. This still
left a life expectancy gap of 5.4 years, however.
Julia Marcus of Kaiser, presenting, commented: “In addition
to timely ART initiation, risk-reduction strategies such as smoking cessation
may further narrow the survival gap." She said they would consider looking at other factors that might impact on life expectancy such as depression, which is more common in people with HIV.
She added: “Future studies should determine if this survival gap persists
in more recent years, and if so, identify factors that may contribute.”