Death rates and hospitalization rates for people living with HIV have fallen dramatically since the introduction of highly active antiretroviral therapy (HAART) a decade ago. Previous studies have shown that death rates have decreased 50% since the introduction of HAART.¹ But how long can people with HIV expect to live, and what are the factors that -- despite the successes of HAART -- continue to keep life expectancy lower for HIV-infected people than HIV-uninfected people? Several epidemiological studies presented at the XVI International AIDS Conference explored these questions.

One such study was a report of data from the Danish HIV Cohort study,² an important cohort that follows all of the people known to be receiving, or to have received, treatment for HIV in Denmark. Each of these nearly 4,000 HIV-infected patients were followed from 1995 to 2005 and compared to approximately 400,000 age- and gender-controlled, HIV-uninfected patients.

The results were encouraging, particularly considering how bleak the prognosis was for HIV-infected people a short 10 years ago. According to the Danish study, someone who is newly HIV infected at age 25 is now expected to live 57.5 years -- or 63.9 years if the person is not coinfected with hepatitis C. However, while this is a large increase in lifespan over the pre-HAART era, life expectancy for HIV-infected individuals still lags behind their HIV-uninfected counterparts: at age 25, people without HIV in Denmark are expected to live 76.2 years.

Graph by Nicolai Lohse, M.D.; reprinted with permission. Click here to download the complete poster.

As life expectancy shifts for HIV-infected people, mortality causes have shifted as well. Earlier reports have indicated that the causes of death in HIV-infected individuals are increasingly related to cardiac and liver problems, with the latter occurring particularly in hepatitis-coinfected individuals. A new analysis of AIDS-related deaths in the United Kingdom (UK) between 1993 and 2004 was presented at this conference,³ in which researchers from the UK's Health Protection Agency/Centre for Infection compared data from the pre- and post-HAART eras. They found that four complications in particular have been increasingly cited as causes of death in HIV-infected adults: pneumocystis pneumonia (6% in the pre-HAART era versus 11% in the post-HAART era), non-Hodgkin's lymphoma (5% versus 9%), cardiovascular disease (5% versus 8%) and liver disease (3% versus 7%).

(The UK study also noted a significant demographic shift in deaths among HIV-infected people: The largest population of HIV-infected people in the UK, men who have sex with men (MSM), now makes up 46% (down from 70%) of the deaths, which are increasingly seen among heterosexual, black Africans. Worth noting as well is the finding that three quarters of UK AIDS deaths occurred in people who were diagnosed late in the course of their disease.)

However, one of the limitations with studies evaluating causes of death is the accuracy of the data. A separate British study presented at this conference,⁴ which looked at autopsies in HIV-infected patients, showed that 71% of all diagnoses and 51% of all primary causes of death were changed following autopsy findings. Alarmingly, diagnoses of opportunistic infections (including Kaposi's sarcoma, tuberculosis and mycobacterium avium complex) were missed in multiple patients before autopsy. Accurately diagnosing cause of death is essential if we are to make the appropriate adjustments in our care of those living with HIV.

Supporting evidence for these changes in cause of death is found in similar trends for cause of hospitalization among HIV-infected patients. Data on 3,863 HIV-infected patients in the French ANRS CO3 Aquitaine Cohort study, which were presented at this conference,⁵ showed a 46% decrease in hospitalizations for HIV-infected patients between 2000 and 2004. During this period, 21% of hospitalizations were caused by bacterial infections, 20% by AIDS events (with a reduction from 24% in 2000 to 11% in 2004, dropping them from the #1 cause of hospitalization to #2), 10% by psychiatric disorders, 9% by cardiovascular events (with an increase from 5% in 2000 to 15% in 2004), 7% by hepatic/gastro-intestinal disorders, 6% by other viral infections and 5% by non-AIDS-associated cancers. Of these causes, all but cardiovascular events and non-AIDS-associated cancers exhibited a statistically significant decline in incidence rate from 2000 to 2004 (see chart below).

Graph by Fabrice Bonnet, M.D.; reprinted with permission. Click here to download the complete poster.

It is unclear how much of the changes in hospitalization rate and causes are related to changes in demographics alone (e.g., increasing age, increasing number of substance users, increasing number of patients with hepatitis coinfection) versus other etiologies (e.g., long-term HIV drug toxicity).

It is this shift in the demographics of HIV in the United States and Europe that highlights an often-underappreciated fact that may impact morbidity and mortality: Unfortunately, not all people with HIV are benefiting equally from HAART. Previous studies have shown fewer life-extending benefits to HAART for older patients, patients who use non-injection drugs, patients with hepatitis coinfection and even for those with higher rates of unemployment or a lack of stable relationships. The Danish HIV Cohort study presented at this conference confirmed increased mortality among older HIV-infected patients and those coinfected with hepatitis C. In addition, a study presented by Melissa R. Pfeiffer and colleagues from the New York City Department of Health and Mental Hygiene looked at New York City AIDS deaths from 1999 to 2003 in MSM and injection drug users.⁶ A significant excess of deaths was seen in the injection drug users, particularly with regard to overdose (21.5% versus 4.1% in MSM) and liver disease (16.4% versus 7.6% in MSM).

Finally, Timothy Lahey, of Dartmouth Medical School, and colleagues reported at this conference on a 636-patient cohort from the Dartmouth-Hitchcock HIV Program comparing HIV-infected rural (n=323) and urban (n=313) patients.⁷ Mortality was shown to be higher in rural versus urban patients (10.4% versus 6.0%, P = .028). Even when controlled for multiple variables (age, sex, race, HIV risk factors, year of diagnosis, travel time, lack of insurance, antiretroviral treatment and pneumocystis pneumonia prophylaxis), the higher mortality rates persisted in rural HIV-infected patients, as shown in the graph below.

Graph by Timothy Lahey, M.D.; reprinted with permission. Click here to download the complete poster.

In fact, rural patients also tended to be more likely to die than urban patients who had similar CD4+ cell counts at first presentation to the program, as illustrated in this graph:

Graph by Timothy Lahey, M.D.; reprinted with permission. Click here to download the complete poster.

The aggregate of these study results in the United States and Europe presents a mixed view of the current state of the epidemic. While life-extending benefits from effective HIV treatments continue to be seen, those living with HIV still have a shorter life expectancy compared with their HIV-uninfected counterparts. Increasing focus is needed on the remaining causes of HIV-related morbidity and mortality so that we can continue to make strides in lengthening our patients' life spans. If the epidemic is increasingly affecting those who are coinfected with hepatitis or who use drugs, more studies and resources are needed to focus on reducing the added risks caused by these conditions. If liver disease, cardiac disease and cancers are becoming more common causes of death, we need to sort out how much of this is related to the long-term effects of HIV and its comorbidities versus the long-term effects of HIV medications. And finally, as the epidemic spreads from major cities to more areas in places like the rural United States, we need to confirm if, and why, rural HIV-infected patients really have increased mortality risks. The same attention must be paid to the growing number of HIV-infected people who must deal with the complications of aging alongside those of HIV. Only through more intense and specific focus on these new disparities will we be able to bridge the gap in life span between those living with and without HIV.

Footnotes

1. U.S. Centers for Disease Control and Prevention. AIDS Surveillance -- Trends 1985-2004. Available at: www.cdc.gov/hiv/topics/surveillance/resources/ slides/trends/slides/AIDS_Trends.pdf. Accessed August 17, 2006.

2. Lohse N, Hansen A-BE, Pedersen G, et al, Danish HIV Cohort Study. Median survival and age-specific mortality of Danish HIV-infected individuals: a comparison with the general population. In: Program and abstracts of the XVI International AIDS Conference; August 13-18, 2006; Toronto, Canada. Abstract MOPE0310.
   View poster: Download PDF

3. Ciancio BC, Forde J, Dougan S, Chadborn T, Delpech V. Trends in mortality and causes of death among HIV-infected individuals diagnosed in the UK: 1993-2004. In: Program and abstracts of the XVI International AIDS Conference; August 13-18, 2006; Toronto, Canada. Abstract THAC0202.

4. Cohen D, Beadsworth M, Jenkins N, Ratcliffe L, Taylor B, Beeching N. Autopsies in HIV: still identifying missed diagnoses. In: Program and abstracts of the XVI International AIDS Conference; August 13-18, 2006; Toronto, Canada. Abstract MOPE0064.
   View poster: Download PowerPoint

5. Bonnet F, Chêne G, Lawson-Ayayi S, et al, Groupe d'Epidemiologie Clinique du SIDA en Aquitaine. Causes of severe morbidity in HIV-infected patients. Aquitaine cohort 2000-2004: the importance of bacterial infections, cardio-vascular, digestive, and psychiatric morbidity. In: Program and abstracts of the XVI International AIDS Conference; August 13-18, 2006; Toronto, Canada. Abstract MOPDB02.
   View poster: Download PowerPoint

6. Pfeiffer MR, Hanna DB, Begier EM, Sepkowitz KA, Torian LV, Sackoff JE. Persistent contribution of substance abuse to excess mortality among persons with AIDS in New York City, 1999-2003. In: Program and abstracts of the XVI International AIDS Conference; August 13-18, 2006; Toronto, Canada. Abstract THAC0201.
   View slides: Download PDF

7. Lahey T, Marsh B, Curtin J, Wood K, Eccles B, von Reyn CF. Increased mortality in rural patients with HIV in New England. In: Program and abstracts of the XVI International AIDS Conference; August 13-18, 2006; Toronto, Canada. Abstract MOPE0305.
   View poster: Download PowerPoint