HIV disease today looks very different than it did two decades ago when the first cases were described -- at least in developed countries where highly active antiretroviral therapy (HAART) is widely available. Most HIV positive people receiving treatment no longer succumb to opportunistic illnesses (OIs) that take advantage of their ravaged immune systems. Instead, people with HIV now live longer and die from a wide variety of other causes. In other words, a growing number of positive people will die with HIV, but not of HIV.

In the HAART era, HIV positive people and their health-care providers must now think about the cumulative impact of chronic HIV infection and the long-term side effects of treatment. In addition, as effective antiretroviral therapy extends the lives of people with HIV, they become prone to conditions that normally occur with greater frequency as people age (such as heart disease, diabetes, and osteoporosis) and progressive conditions that may take decades to cause significant illness or death (such as chronic viral hepatitis). Given the shifts in the types of conditions now seen in people with HIV, it may be time to rethink the definition of AIDS.

Changing Patterns of Mortality

In the 1980s and early 1990s, HIV/AIDS was a major cause of death among adults in the U.S., with the mortality (death) rate climbing every year from 1987 to 1994. By the latter year, the disease had become the leading cause of death among adults 25-44 years of age. A significant decrease in AIDS mortality first became apparent in 1996, the year after the first protease inhibitor (PI) was introduced. In 1997 the number of HIV/AIDS deaths fell by nearly 50%, followed by a further 20% reduction in 1998. By 1999, however, the decline had leveled off. According to the Centers for Disease Control and Prevention (CDC), HIV/AIDS deaths fell from more than 51,000 in 1995 to about 16,000 in 2002 (the latest year for which data are available).

In 2002 HIV/AIDS was the fifth leading cause of death for adults aged 25-44, but not in the top 15 for the population as a whole. It is important to emphasize, however, that the overall HIV/AIDS mortality rate hides differences among demographic groups. In 1996, for example, while HIV/AIDS deaths dropped by 20% among white people, the rate declined by just 2% among black individuals. The decrease among men who have sex with men was three times as great as that for injection drug users (IDUs), while the mortality rate actually rose for people infected through heterosexual contact. And in 2001 HIV/AIDS remained the leading cause of death for black women aged 25-34, compared with the third leading cause for black men, sixth for white men, and seventh for white women in the same age group.

Looking at the population as a whole, the top three causes of death in 2002 were heart disease, cancer, and stroke, followed by diabetes at number six and chronic liver disease/cirrhosis at number 12. As discussed below, these conditions now account for a growing proportion of deaths among people with HIV as well.

AIDS Then and Now

Perhaps the most evident change in the epidemic over the past 20 years has been the increase in the length of time people with access to good health care can expect to live after testing HIV positive or being diagnosed with AIDS. According to Dennis Osmond, PhD, of the University of California at San Francisco, in the earliest years of the epidemic individuals could expect to live about one year after an AIDS diagnosis. Survival times with AIDS began to increase in the mid-1980s as OI treatment improved. With the wider use of OI prophylaxis (preventive therapy), the time between seroconversion and an AIDS diagnosis increased as well. The introduction of PIs in 1995 -- which enabled the construction of potent combination regimens consisting of drugs from two or more classes -- led to a tremendous expansion of time between seroconversion and progression to AIDS or death. In the Multicenter AIDS Cohort Study (MACS), for example, the estimated time from seroconversion to death for a person infected at age 30 rose to a median of about 13 years by the 1995-1997 period.

Today many individuals have survived -- and even thrived -- with HIV for upwards of two decades. Because HIV/AIDS is still a relatively young disease, it is too soon to know the upper limit of survival for HIV positive people on HAART. While the long-term toxicities of antiretroviral therapy remain an urgent concern, some experts now cautiously predict that HIV positive people who receive optimal care may ultimately live a near-normal lifespan.

Who Gets AIDS Today?

Where HAART is widely available, many people starting treatment in recent years have never reached a CD4 cell nadir (lowest-ever level) below 200 cells/mm³ or developed an AIDS-defining OI. Those who do typically fall into one of the following categories:

• individuals who have been infected with HIV for a long time and may have previously received suboptimal treatment -- including nucleoside reverse transcriptase inhibitor (NRTI) monotherapy -- before the advent of HAART; such people often have HIV that is resistant to many drugs and thus have limited treatment options
• people who have access to effective HAART, but either are not able to tolerate the drugs or fail to achieve adequate adherence
• individuals who cannot afford antiretroviral therapy and are unable to access benefits through programs such as Medicaid or state AIDS Drug Assistance Programs (ADAPs)
• people who have never been tested for HIV and do not seek care until they have already experienced significant immune system decline and perhaps developed AIDS-related symptoms.

The CDC estimates that as many as two-thirds of people living with HIV in the U.S. do not know they are infected, and nearly one-half are diagnosed after their immune function has already markedly declined. Such cases are most often seen among marginalized populations (including racial/ethnic minorities, the homeless, and IDUs) and tend to be concentrated in inner cities. Socioeconomic factors such as poverty, substance use, and coexisting psychiatric illnesses can have a profound impact on access to and success of anti-HIV therapy.

Several studies presented at the XV International AIDS Conference held this past summer in Bangkok looked at factors associated with late testing and treatment. Scott Kellerman, MD, and J. Blair from the CDC found that 1,175 of 2,063 people (57%) seen in Boston, Chicago, Hartford, Los Angeles, and San Francisco received an AIDS diagnosis within just six months of first testing HIV positive. The chances of late HIV detection were greater among African Americans (1.4 times as high as whites), Latinos/Hispanics (1.7 times as high), and individuals over age 40. African Americans were twice as likely as whites to develop AIDS before receiving antiretroviral therapy. A study from North Carolina found that late presentation was linked to male sex and living in a rural area. And an analysis of the French Hospital Database on HIV presented by Murielle Mary-Krause, MD, and colleagues showed late treatment to be associated with older age, HIV acquisition by a route other than male-to-male sex, and immigration from sub-Saharan Africa.

It is hoped that the CDC's recently announced emphasis on routine testing, along with the availability of a new rapid oral HIV test, will reduce the number of people who are unaware that they harbor HIV. Offering regular testing to individuals at risk can help bring more people into treatment earlier in the course of their disease, when antiretroviral therapy can do the most good.

How Late Is Too Late?

Treatment of late-diagnosed individuals can be challenging, since they may require treatment or prophylaxis for OIs in addition to HAART, thereby increasing the risk of additive side effects and drug interactions. In addition, people who start HAART with low CD4 cell counts and/or high HIV viral loads remain at greater risk for progression to an AIDS-defining illness or death compared with those who start therapy with less compromised immune function.

For example, Andrew Phillips and colleagues with the CASCADE Collaboration reported in the January 2, 2004 issue of AIDS that the short-term risk of progression to AIDS rose in a linear fashion as CD4 cell count decreased and as HIV viral load and age increased. While a 25-year-old individual with 500 CD4 cells/mm³ and a viral load of 3,000 copies/mL had a miniscule 0.3% probability of developing AIDS within six months (chosen as a typical interval between clinic visits), the rate climbed to 44.8% for a 55-year-old person with 50 CD4 cells/mm³ and a viral load of 300,000 copies/mL. (For a detailed chart entitled "Predicted 6-month risk of AIDS according to age and current CD4 cell count and viral load," see the October 2004 adult U.S. federal treatment guidelines, table 3b, at http://aidsinfo.nih.gov.)

Data from the ART Cohort Collaboration illustrate the perils of late treatment. Matthias Egger, MD, and colleagues reported findings based on an analysis of some 12,500 participants in a dozen American and European clinical trials in the July 13, 2002 issue of The Lancet, and presented an update at the Bangkok AIDS conference. In this analysis, progression to AIDS or death was strongly associated with viral loads above 100,000 copies/mL and lower baseline CD4 cell counts, with risk increasing as counts fell below 350, 200, 100, and 50 CD4 cells/mm³. The rate of progression among individuals starting anti-HIV treatment with fewer than 50 cells/mm³ was five times as high as the rate among people with more than 350 cells/mm³. Starting therapy with CD4 cell counts above 350 cells/mm³ did not appreciably improve outcomes, however, lending support to the 350 cells/mm³ threshold recommended in the U.S. HIV treatment guidelines. But another analysis of the ART Cohort presented at the same conference by Caroline Sabin showed an increased risk of progression among individuals starting treatment with 200-350 CD4 cells/mm³, and other research suggests that IDUs may need to initiate antiretroviral therapy at higher CD4 counts to derive the same benefit that non-IDUs get from starting at lower CD4 cell levels (see "News Briefs" in this issue), so the question as to whether people within this range should start therapy remains somewhat murky.

Despite abundant evidence that deteriorating immune function and increasing HIV viral load levels are associated with a greater risk of progression to AIDS or death, research on the whole indicates that even people who begin treatment with severe immunosuppression, very low CD4 cell counts, high viral loads, and/or symptomatic AIDS can still derive significant benefit from antiretroviral therapy. For example, S. Koltar and colleagues reported in the November 15, 2004 issue of Clinical Infectious Diseases (CID) that in the ACTG 362 study, which included 612 participants followed for 3-5 years, HIV disease progression was uncommon (1.75 new AIDS-defining conditions per 100 person-years [PY]) in people who started HAART with CD4 cell counts below 50 cells/mm³ -- 62% of whom had previously been diagnosed with at least one AIDS-defining illness -- as long as they were then able to achieve sustained CD4 cell increases of at least 100 cells/mm³ and maintain low HIV viral loads.

AIDS vs. Non-AIDS Mortality

Dying with HIV disease today is not the same as dying of AIDS a decade or two ago. As the number of deaths due to OIs and other AIDS-defining conditions has fallen, the proportion of deaths due to all other causes has consequently risen. Importantly, such a shift does not necessarily indicate an absolute increase in the number of non-AIDS deaths, since a rising proportion of deaths due to one cause may simply reflect a falling proportion of deaths due to another.

One of the best snapshots of the epidemic is the HIV Outpatient Study (HOPS), which monitors more than 5,500 participants at a dozen U.S. HIV clinics. A HOPS analysis published in the March 26, 1998 issue of the New England Journal of Medicine (NEJM) was the first to show a marked decline in mortality at the dawn of the HAART era. Among the 1,255 subjects then in the database, the overall death rate decreased from 29.4 per 100 PY in 1995 to 8.8 per 100 PY in mid-1997. According to an update presented by Frank Palella, MD, at the 11th Conference on Retroviruses and Opportunistic Infections in February 2004, the mortality rate was 2.2 per 100 PY in 2002 -- a figure that has been roughly stable since 1998. Between 1996 and 2002 the proportion of non-OI deaths rose from 46% to 72%. "If someone takes [HAART], they will live longer and when death occurs, it will not be due to an AIDS-related condition," Dr. Palella concluded.

Looking at a large European cohort, Amanda Mocroft, PhD, and colleagues reported in the August 16, 2002 issue of AIDS that in the EuroSIDA cohort, which includes more than 8,500 HIV positive participants, the overall death rate fell from 15.6 to 2.7 per 100 PY between 1994 and 2001. In addition, the proportion of AIDS deaths decreased by 23%, while the proportion of non-AIDS deaths rose by 32% during the same period. Similarly, Ard van Sighem and colleagues with the Dutch ATHENA study reported in the October 17, 2003 issue of the same journal that among more than 3,700 participants using HAART, HIV-related mortality decreased from 3.8 to 0.7 per 100 PY between 1996 and 2000, while non-HIV-related mortality remained constant. (This research team defined deaths related to antiretroviral therapy as non-HIV-related.)

Finally, in the October 18, 2003 issue of The Lancet, Kholoud Porter, MD, and colleagues, also with the CASCADE Collaboration, reported on a study of 7,740 participants from 22 cohorts in Europe, Australia, and Canada, categorized into three groups based on when they seroconverted: pre-1997 (when HAART was introduced in these countries), 1997-1998 (limited HAART), or 1999-2001 (widespread HAART). By 1997 the rate of AIDS-related deaths had decreased by about 50%, and by 2001 it had fallen by 80%.

Shifting Causes of Illness and Death

A wide variety of conditions fall within the broad category of "non-OI" or "non-AIDS-related" causes of morbidity (illness) and mortality. These include age-related conditions such as cardiovascular disease and diabetes, as well as toxicities associated with antiretroviral therapy.

In Dr. Koltar's ACTG 362 cohort, five participants died of AIDS-related causes between 1997 and 2002, while five died of cardiovascular conditions and five due to liver failure related to hepatitis B or C virus (HBV and HCV, respectively). In Dr. Palella's HOPS cohort, deaths due to three major OIs -- Pneumocystis carinii (now P. jiroveci) pneumonia (PCP), Mycobacterium avium complex (MAC), and cytomegalovirus (CMV) -- declined from 21.9 per 100 PY to 3.7 per 100 PY between 1994 and mid-1997. In the 2000-2002 period the most common non-OI causes of death were liver (36%), lung (23%), cardiovascular (17%), and kidney (10%) conditions.

In a different type of analysis, Richard Selik, MD, of the CDC and colleagues examined death certificate data for all deaths in the U.S. between 1987 and 1999; results were published in the April 1, 2002 issue of the Journal of Acquired Immune Deficiency Syndromes (JAIDS). From 1995 to 1999, rates of several AIDS-defining illnesses declined, including Kaposi's sarcoma (5% to 3%), CMV (7% to 3%), HIV encephalopathy (6% to 4%), and wasting (10% to 7%). At the same time, rates of deaths due to certain non-AIDS causes rose: septicemia (blood poisoning; 9% to 13%), liver disease (5% to 12%), kidney disease (6% to 9%), and heart disease (4% to 7%). The researchers suggested that improved OI prophylaxis and treatment likely contributed to decreased OI rates, while toxicity associated with antiretroviral therapy may have contributed to higher rates of liver, kidney, and heart problems. At the 2004 Retrovirus conference Dr. Selik presented 1987-1999 death certificate data for San Francisco and New York City -- two early epicenters of the epidemic -- showing that non-AIDS deaths rose from 11% to 23% during the study period.

While most studies have shown a decline in AIDS-related mortality, proportions and absolute numbers of deaths vary considerably by locale and demographic group. At the 2001 Retrovirus conference, S. Ahmad from Chicago's Cook County Hospital presented data from a retrospective chart review of all HIV inpatient deaths between January 1998 and September 2000. This patient population was predominantly comprised of substance-using African American men. Among those who died, more than half were not receiving antiretroviral therapy and about half had CD4 cell counts below 50 cells/mm³. It is therefore not surprising that OIs remained common in this cohort, accounting for about 60% of all deaths. These results underline the importance of improving access to care and promoting early testing and treatment, especially among marginalized populations.

As the relative proportion of OI or AIDS-related deaths among HIV positive people has decreased, certain non-AIDS-defining conditions have received particular attention as important new threats for people living with HIV. Liver problems are sometimes related to antiretroviral drug toxicity, but liver damage due to chronic viral hepatitis is a growing concern as people coinfected with HIV plus HBV and/or HCV live longer thanks to HAART. But HAART itself can lead to new problems, as exemplified by the potential for increased risk of cardiovascular disease due to blood lipid (fat) and glucose (sugar) elevations associated with PI-based therapy. Trends in AIDS-defining and non-AIDS-related malignancies (cancers) are less clear, but most research indicates that HIV positive individuals face a higher cancer risk than their HIV negative counterparts. Liver problems, heart disease, and malignancies in people with HIV will be discussed in detail in the following sections.

Liver Problems

Liver problems in people with HIV tend to fall into two categories: progressive liver damage due to chronic viral hepatitis coinfection, and hepatotoxicity (liver toxicity) associated with certain antiretroviral drugs.

In the HAART era liver disease has become a major cause of hospitalization and death among people with HIV. In the February 1, 2001 issue of CID, for example, Iona Bica, MD, and colleagues reported that during the 1998-1999 period, 50% of deaths (11 of 22) of HIV positive people at Boston's New England Medical Center were due to end-stage liver disease (ESLD), compared with 12% in 1991. In Dr. Ahmad's study of patients at Cook County Hospital (about half of whom were coinfected with HCV), ESLD was the second leading cause of mortality, accounting for 35% of deaths. And in the Women's Interagency HIV Study (WIHS), liver disease was the leading non-AIDS cause of mortality in the 1994-2000 period, accounting for about 20% of all deaths.

A similar situation exists in Europe, especially in Spain and Italy. At the Instituto de Salud Carlos III in Madrid, ESLD was responsible for 43% of deaths of people with HIV in 2000, about one-third of whom were coinfected with HCV. In the French Aquitaine cohort, HCV-related ESLD was the number one cause of mortality in the 1998-1999 period (accounting for 29% of deaths), while in the EuroSIDA cohort liver-related mortality has been a leading cause of death since 2000.

Chronic Viral Hepatitis

About one-quarter of people with chronic hepatitis B or C will experience progression to liver fibrosis (buildup of fibrous tissue), cirrhosis (scarring), and/or hepatocellular carcinoma (a type of liver cancer), a process that typically takes 10-40 years. Before the advent of HAART, coinfected individuals and those with HIV alone fared similarly, because most died of AIDS-related causes before ESLD developed. Today this picture has changed. As reported in the October 21, 2004 issue of AIDS, Maurizio Bonacini, MD, and colleagues reported that HIV positive individuals coinfected with both HBV and HCV were twice as likely to die of a liver-related cause as those with either HIV/HBV or HIV/HCV, who in turn had about twice the risk as people with HIV alone (28%, 13-15%, and 6%, respectively).

Some HIV positive people coinfected with HCV and/or chronic active HBV are more likely to progress to serious liver disease, and to progress more rapidly, than those with viral hepatitis alone. (For more on this topic, see "HIV and Hepatitis Coinfection," BETA, Winter 2003.) At a June 2002 National Institutes of Health (NIH) consensus conference on management of hepatitis C, David Thomas, MD, cited a meta-analysis showing that HIV/HCV-coinfected people had a two-fold greater risk of cirrhosis and a six-fold greater risk of ESLD than those with HCV alone. Abdul Mohsen and colleagues from King's College in London estimated that the average time from HCV infection to the onset of cirrhosis was 23 years in coinfected people, compared with 32 years in people with HCV alone.

However, much early research on coinfection was conducted before the advent of HAART. More recent studies suggest that accelerated liver disease progression may be a consequence of compromised immune function, and that HIV positive people with well-controlled HIV disease may do as well as their HIV negative counterparts. Based on a case-control study of 116 HIV/HCV-coinfected subjects and 235 individuals with hepatitis C alone, for example, Eugenia Mariné-Barjoan and colleagues reported in the November 5, 2004 issue of AIDS that liver fibrosis progressed significantly more slowly in people who had been on HAART longer and in those who had a shorter interval between presumed HCV infection and initiation of anti-HIV therapy.

Research is less clear about the impact of hepatitis C on HIV disease. A majority of studies indicate that it does not have a detrimental effect. For example, Ellen Tedaldi, MD, and colleagues reported in the February 1, 2004 CID that among a cohort of 823 HIV positive individuals, those who were coinfected with HCV did not have a higher rate of progression to AIDS than those with HIV alone.

Other research supports the opposite conclusion. In the Swiss HIV Cohort of more than 3,000 participants, HIV/HCV-coinfected individuals were more likely to develop OIs (8% vs. 5%) and more than twice as likely to die of any cause (9% vs. 4%) than those with HIV alone.

Studies more consistently show that HIV/HCV-coinfected individuals experience slower immune recovery after starting antiretroviral therapy. For example, J. Martin, MD, and colleagues determined that after two years on HAART, CD4 cell counts increased by an average of 53 cells/mm³ in coinfected people compared with 111 cells/mm³ in those with HIV alone.

Fortunately, treatment for hepatitis C has improved in recent years with the adoption of pegylated interferon (Pegasys or