Why People With HIV Still Die -- And Why They Don't Have To ...

• Introduction
• What Kills People With HIV Today?
• Hepatitis Outranks Other Death Threats
• Can Liver Deaths Be Prevented?
• Bad Adherence Can Kill
• Fateful Impact of a Late Diagnosis
• AIDS Cancers and Non-AIDS Cancers
• How Big Is the Threat From Heart Disease?
• Dabbling in HIV Medicine, Dabbling in Death
• ... And All Those Other Reasons
• Acknowledgments
• References and Notes

Introduction

In countries with good access to antiretrovirals, people still die from AIDS if it is diagnosed too late, if they have been unable to take antiretrovirals (for example, because of intolerable side effects), or if they have a condition that antiretrovirals may not improve, such as progressive multifocal leukoencephalopathy. People with HIV infection who respond well to antiretrovirals may die from other causes such as liver disease, heart disease, cancer, or accidental deaths. The risk of death from many of these conditions can be reduced by behavior change. This article describes recent research on HIV-related mortality in developed countries, relevant case reports, and steps clinicians might take to lower the death rate among people with HIV infection.

Clichés cling to medical literature as insistently as burrs to burlap. But no cliché has clung to HIV literature as tenaciously as one that arose circa 1997, when AIDS mortality began its plunge in North America, Western Europe, and Australia. Every other published article and every third meeting abstract -- it sometimes seems -- starts with some version of this trusty formula:

The advent of potent antiretroviral therapy has led to dramatic declines in deaths from AIDS. However ...

The study's title always gives away the next sentence: "However, treatment-induced toxicities have grown more worrisome as people live longer," or "However, failure to eradicate HIV has increased the challenge to develop adjunctive immune-based therapies," or "However, the prospect of life-long therapy has prompted intense study of treatment interruptions."

So far, apparently, no one has penned an equally defensible corollary to this ubiquitous bromide: "However, an awful lot of people still die with HIV." One need not even reference the thousands who die daily around the globe because they cannot get antiretrovirals.

This article does not address those untreated throngs, or people in the United States, for example, who die killing time on an AIDS Drug Assistance Program (ADAP) waiting list.¹

Instead, this IAPAC Monthly article considers the thousands who can and usually do get potent antiretrovirals but die anyway. Their numbers may be dwindling, but they are far from small. The US Centers for Disease Control and Prevention (CDC) estimates that 50,610 people with HIV died in the United States in 1995, the year before potent combinations turned the tide.² By 1998 better treatment had more than halved that number, but 19,005 people with HIV still died in the United States.³ The tally continued to drop in the 21st century but stood at a still dismaying 16,371 in 2002.³

EuroSIDA analyzed all 1,826 deaths in the pan-European cohort from 1994 through 2001.⁴ The largest proportion, 43.8 percent, died in the pre-HAART years 1994 and 1995, but one third died in 1996-1997, and nearly one quarter died in 1998-2001. At the Owen Clinic of the University of California, San Diego, a study of nearly 5,000 patients who sought care for HIV showed a drop in the death rate starting in 1995, but then an upswing among those who came to the clinic after 1998.⁵ A survey of almost 3,000 people enrolled in five US antiretroviral trials between December 1996 and December 2001 plotted a dogged doubling of the death rate every 12 months, from 3.9 percent after one year of follow-up, to 7.9 percent after two, and to 13.1 percent after three.⁶

Should so many people with HIV still be dying? Can all their deaths be ascribed to HIV's remorseless immune carnage, its maddening knack to morph toward resistance? No, a growing stack of careful studies shows. E-mail surveys and interviews with more than two dozen HIV clinicians and researchers across the United States, Europe, and elsewhere confirm that other forces are at play. A few examples:

• At a major teaching hospital in Texas, Pneumocystis carinii pneumonia (PCP) accounted for an equivalent proportion of deaths before HAART in 1995 (21 of 112, or 19 percent) and well into the HAART era in 1999-2000 (15 of 88, or 17 percent, P=0.76).⁷

• An analysis of 66 deaths in France's Aquitaine cohort in 1998 and 1999 blamed 11 of them (17 percent) on treatment-induced toxicities.⁸

• In British Columbia, where access to antiretrovirals is universal and free, only pretreatment CD4 count and intermittent therapy predicted death in a study of 1,282 people beginning their first antiretrovirals between August 1996 and December 1999.⁹

Even the most skilled HIV clinicians working in an ideal healthcare system cannot prevent all deaths like these. But surely different decisions in such cases -- by physicians or by infected people -- would have saved lives. And certainly treatment by clinicians not skilled in HIV medicine has ruined many a person's shot at successful therapy -- a fact documented before^10, 11 and after¹² potent therapies arrived.

Through a review of the literature and input from top HIV treaters and researchers, this article will try to: (1) itemize the emerging and persisting causes of death in HIV-infected people with access to good treatment, and (2) pinpoint those causes that may be avoidable. This second goal may seem overly ambitious to some readers, while others will judge it likely to yield only painfully obvious answers.

The research behind this article suggests that many avoidable causes of death are painfully obvious, but no less painful, therefore, to the people who die. Some avoidable causes are more surprising. Yet the ultimate aim of this article is not to serve up investigative coups, but to catalog the evolving causes of death in antiretroviral-rich countries at this point in the epidemic as a handy reminder for HIV clinicians and the people whom they treat.

What Kills People With HIV Today?

Because potent antiretrovirals can keep CD4 counts above the danger zone where opportunists operate, deaths with classic AIDS diseases such as PCP, Kaposi's sarcoma (KS), and cytomegalovirus (CMV) infection have plunged in adequately treated people. This pervasive trend fosters the notion -- confirmed in some small cohorts -- that non-AIDS diseases have caught up with (and sometimes surpassed) standard AIDS-defining diseases as causes of death. For example:

• A study of 255 deaths of people with HIV from 1995 through 1999 at the University Hospitals of Cleveland found an AIDS-defining cause of death in 60 percent in 1995 versus 31 percent in 1999 (P<0.001).¹³ For the same two years, the proportion of deaths unrelated to HIV rose from 5 percent to 19 percent.

• At Chicago's Cook County Hospital, where the HIV population is 73 percent African American, 14 percent Latino, and 33 percent female, death certificates cited non-HIV causes in 13 percent of men and 21 percent of women in 1996 compared with 31 percent of men and 33 percent of women in 1999.¹⁴ HIV-related deaths dropped accordingly.

• At Parkland Memorial Hospital in Dallas, a comparison of deaths of 210 inpatients with HIV in 1995 and 1999-2000 charted a drop in "HIV-associated illness" as the cause from 62 to 47 percent (P=0.03) and a drop in AIDS-defining illness from 51 to 38 percent (P=0.058).⁷ Meanwhile, the proportion of non-AIDS-related causes of death climbed from 34 to 48 percent (P=0.048).

But other recent cohort studies show little slippage in proportions of people dying because of AIDS definers:

• Among 422 deaths of HIV-infected people in France during 2000, 51 percent involved at least one AIDS-defining illness, most often wasting, non-Hodgkin lymphoma (NHL), CMV, and PCP.¹⁵

• Analysis of 204 deaths of people attending an Atlanta clinic showed a significant drop in mortality from 1995 to 2000, but no significant change in the main causes of death -- infection, organ failure, neoplasia, and end-stage AIDS.¹⁶

So what's happening? Are non-AIDS deaths replacing AIDS deaths or not? IAPAC Monthly's unscientific sampling of HIV clinicians and clinical researchers suggests one answer. AIDS-defining causes of death remain plentiful today for three reasons, this survey indicates (Table 1):

1. Because many people -- often substance abusers and people with mental illness -- are unable to stick with an antiretroviral regimen, they never stop HIV's assault on the immune system and die with classic AIDS diseases.

2. A large proportion of people dying with HIV today go undiagnosed until they have advanced, unsalvageable AIDS.

3. People still die from NHL, an AIDS-defining disease.

Along with those AIDS-defining causes, liver disease -- usually kindled by hepatitis C virus (HCV) coinfection -- kills a substantial proportion of people with HIV. In an interview with IAPAC Monthly, Daniel Kuritzkes (Brigham and Women's Hospital, Boston) explained that most people who die with AIDS despite antiretroviral access "are coming into treatment late or are unable to comply with therapy. The most challenging patients -- usually active substance abusers -- are in and out of care, on and off therapy, and keep coming in with recurrent Pneumocystis, toxo, cryptococcal meningitis, those kinds of things. They probably would do well if they could get on therapy and stay on it.

"A much smaller group in that same cause category consists of people who have highly resistant virus and are slowly progressing and dying of AIDS.

"The other big category includes the people dying because of liver disease, usually with hepatitis C."

In an e-mail note, W. Christopher Mathews (University of California, San Diego) added another contributor to the sustained death rate from AIDS definers -- age. People "over 55 or so," he observed, tolerate classic opportunistic infections more poorly than younger people, especially when the diagnosis comes late.

Three large studies -- two involving cohorts and one analyzing death certificates -- supply some statistics that flesh out these clinical impressions. Scrutinizing deaths of 1,826 EuroSIDA cohort members in 1994-1995 ("pre-HAART"), 1996-1997 ("early HAART"), and 1998 or later ("late HAART"), researchers found that the proportion of deaths from non-HIV causes rose from 22.6 percent in 1994 to 51.6 percent in 2000 and later (P<0.0001).⁴ In the same two periods, the proportion of HIV-related deaths slumped from 54.0 percent to 16.7 percent (P<0.0001). But the incidence of deaths before an AIDS diagnosis in the cohort dropped from a low 3.1 per 100 person-years in 1995 to 1.1 per 100 person years in 2000-2001. The relative proportion of these non-AIDS deaths grew only because the overall incidence of death plunged from 21.6 per 100 person-years in 1995 to 2.7 in 2000-2001.

The CASCADE Collaboration of 20 European and Australian HIV cohorts gauged the effect of seroconversion year on the risk of an AIDS diagnosis or death without AIDS in 6,941 people with known seroconversion dates from 1979 through 2001.¹⁷ To analyze these changes the CASCADE team used a "competing risks model" that allows for the fact that a rising incidence of one AIDS disease may simply reflect a falling incidence of another disease. Compared with 1994-1996, the researchers traced a significant reduction in the incidence of all major AIDS diagnoses, but not in death without AIDS or death from minor AIDS-defining illnesses (see note 18).

The CASCADE authors argue that some studies may fail to spot drops in particular AIDS diagnoses because they do not use a competing risks model or because they include too few diagnoses to identify significant changes. But the CASCADE study addresses only first AIDS diagnoses -- not which AIDS diagnoses may be the proximate causes of death. For example, the authors observe, they found a significant drop in lymphoma as a first AIDS diagnosis after 1997, but that could mean lymphoma arises more often as a later AIDS diagnosis. So if people now survive an AIDS-defining disease like esophageal candidiasis but succumb to lymphoma, the cancer may kill them.

Notably, the one type of death measured in the CASCADE study -- death without AIDS -- did not drop during the four seroconversion periods analyzed:

• 1979-1990: 9.4 deaths without AIDS per 1,000 person-years

• 1991-1993: 10.0 deaths without AIDS per 1,000 person-years

• 1994-1996: 13.4 deaths without AIDS per 1,000 person-years

• 1997-2001: 11.4 deaths without AIDS per 1,000 person-years

In a multivariate analysis adjusting for gender, age, and HIV exposure category, the risk of death without AIDS did not vary significantly between the 1994-1996 seroconversion group and the 1997-2001 group. In other words, across these 20 European and Australian cohorts, the rate of non-AIDS causes of death stayed constant through 2001.

A different kind of study, one analyzing thousands of death certificates of HIV-infected people dying in the United States from 1987 through 1999, did find a jump in the proportion of deaths with an underlying cause unrelated to HIV.¹⁹ But after 1995, the CDC charted sharp drops in the absolute number of deaths from: (1) HIV infection as the "underlying cause," (2) AIDS-defining disease, and (3) "other underlying causes." There were four exceptions to declines in deaths from other underlying causes:

• Hepatitis B: from 226 deaths in 1997 to 324 in 1999

• Hepatitis non-A/non-B (including C): from 476 deaths in 1997 to 807 in 1999

• Liver cancer: from 33 deaths in 1997 to 47 in 1999

• Other liver disease: from 1,330 in 1997 to 1,432 in 1999

Hepatitis Outranks Other Death Threats

The CDC's finding of surging liver-related mortality in people with HIV¹⁹ would not surprise the 29 clinicians IAPAC Monthly polled on leading causes of death today. The survey included a long list of causes culled from recent literature on HIV mortality and presented in random order. Respondents were asked to check off which causes "have proved most important in your experience." Two physicians marked every cause, while most picked four or five. Twenty-five of the 29 checked off "liver complications" and usually specified HCV infection. Table 1 ranks the results.

The HIV literature is rife with reports on liver problems -- often hepatitis coinfection -- and mortality. Most studies saw a surge in the proportion of liver-related deaths since the dawn of potent antiretroviral therapy, though there are exceptions. A chart review of 262 University of Basel inpatients found no jump in the rate of liver failure as the cause of death from pre-HAART days (1994-1996) to the HAART era (1997-1999).²⁰ But a much larger Swiss HIV Cohort Study determined that HCV coinfection greatly raises the risk of death in people successfully treated for HIV.²¹ This analysis involved 2,318 HCV-negative people and 1,645 with HIV plus HCV. The Swiss team found that the "excess death rate" (compared with the general population) in people without HCV and successfully treated HIV was lower than the excess death rate in successfully treated cancer patients without HIV. But the death rate among people with HCV and successfully treated HIV jumped above the rate in successfully treated cancer patients.

In the IAPAC Monthly survey, Swiss HIV Cohort Study investigator Bernard Hirschel (University Hospital Geneva) rated the listed causes of death from 1 (least frequent) to 5 (most frequent). He assigned a 5 only to liver complications "linked to hepatitis C and intravenous drug use." He gave a 3 to "antiretroviral interactions compromising antiviral efficacy," explaining that "patients with end-stage liver failure are especially difficult to treat." At London's Chelsea and Westminster Hospital, Graeme Moyle estimated that liver problems account for 40 percent of deaths in people with HIV -- the leading cause by far.

At a few centers in the United States, HCV infection and other hepatic problems have not emerged as a leading killer. University Hospitals of Cleveland researchers found that liver failure caused the deaths of only 8.3 percent of those dying in 2001, and they saw no rise in liver-related deaths over recent years.²² At Parkland Memorial Hospital in Dallas, mortality from end-stage liver disease did not change from 1995 to 1999-2000.⁷ And from the Owen Clinic in San Diego, Mathews wrote to IAPAC Monthly that, "despite all the reports of HCV-related mortality, we haven't seen as much as might be expected given a 24 percent overall coinfection rate (70 percent among our injecting drug users with HIV)."

But those are the exceptions. Bigger studies consistently show jumps in the proportion of HIV-infected people dying with hepatitis and other liver problems:

• In a EuroSIDA study of 1,826 deaths from 1994 to 2001, the rate of liver-related deaths fell from 6.85 per 1,000 person-years in 1994 to 2.73 in 1998 but then began climbing again, to 3.45 in 2000 and later.⁴ Since January 2000, hepatitis emerged as the leading killer in EuroSIDA, accounting for 17 percent of those who died.

• A survey of US death certificates of people dying with HIV from 1987 through 1999 saw decided gains in the proportions dying with hepatitis B or hepatitis non-A/non-B (including C) starting in 1996.¹⁹

• A Women's Interagency HIV Study (WIHS) involving 2,059 infected women monitored from 1994 to 2000 rated "hepatic disease" the leading non-AIDS cause of death, in 21 percent of women who died.²³

• A six-city CDC survey of 7,188 deaths of people with HIV found a 70 percent jump in the risk of death from liver disease when comparing the pre-HAART era (1992-1995) with the HAART era (1996-2000).²⁴

Studies relying on death certificates to establish the cause of death may underestimate the rate of liver-related deaths, according to Barbara McGovern and colleagues at Boston's New England Medical Center.²⁵ When working on their small study of increasing mortality from end-stage liver disease,²⁶ they were "surprised to find death certificates that reported 'acquired immunodeficiency syndrome' as the immediate cause of death when chart review clearly showed that the patient died of complications of end-stage liver disease."²⁵

Can Liver Deaths Be Prevented?

Certain deaths from hepatitis and its complications can surely be prevented or at least delayed -- sometimes rather simply, but often only with great effort. Among people coinfected with HCV, hepatitis therapy is suboptimal for many. "Most of them have genotype 1a," Kuritzkes explained, "so they are not responding well to interferon and ribavirin therapy." In a 106-person Spanish trial of interferon plus ribavirin for HCV infection in people with HIV, study participants with genotype 2 or 3 had HCV response rates seven times higher than people with genotype 1 or 4.²⁷

Still, everyone with HIV infection should be tested for HCV, the US Public Health Service recommends,²⁸ because knowing a person's HCV status is critical to treating HIV infection. The guidelines also recommend vaccinating HCV-coinfected people against hepatitis A virus (HAV) and hepatitis B virus (HBV) (see note 29). These experts call for the HBV vaccine in all children with HIV (see Table 10 in reference 28). Many clinicians believe sexually active or drug-using people with HIV -- even those without HCV -- should get the HAV and HBV vaccines to protect them from these bloodborne viruses.

Joel Gallant (Johns Hopkins University, Baltimore) told the story of "two recent patients -- both gay men -- whose doctors never vaccinated them against HBV, despite the clear indications. They became infected both with HIV and HBV, developed chronic hepatitis, and ultimately died, one of cirrhosis and the other of hepatocellular carcinoma -- tragically preventable deaths. I should add that both were doing very well from the standpoint of their HIV infection when they died."

As these two cases illustrate, people carrying both HIV and a hepatitis virus often respond well to antiretroviral therapy. In one US series of 162 people with HIV and HBV or HCV, 88 percent tolerated their antiretrovirals well with little evidence of toxicity on liver function tests.³⁰ And potent antiretroviral therapy appears to lower the risk of liver-related death far more than it raises the risk of deadly liver toxicity. A study of 285 HIV/HCV-coinfected people at the University of Bonn figured the following liver-related death rates in three antiretroviral treatment groups:³¹

• HAART: 0.45 deaths per 100 person-years

• Nucleosides only: 0.69 deaths per 100 person-years

• No antiretrovirals: 1.70 deaths per 100 person years

Five people (9 percent) taking only nucleosides and 13 taking HAART (14 percent) had severe drug-related liver toxicity, but no one died as a result. The survival benefit with HAART, the authors conclude, "seems to outweigh by far the associated risks of severe hepatotoxicity."³¹

Results like these inspired several IAPAC Monthly survey respondents to urge an assertive approach to both diagnosis and treatment of HCV infection in people with HIV. "We need to be more aggressive with treatment of HIV-HCV," wrote José Arribas (Hospital La Paz, Madrid). "Even if the efficacy is less than in HCV-monoinfected people, treatment might be cost-effective."

Yet the risks are there. Nevirapine's (NVP) label carries a "black box" warning about fatal hepatotoxicity. Efavirenz (EFV), stavudine (d4T),³² and the protease inhibitors (PIs) -- especially ritonavir (RTV)-boosted PIs³⁰ -- can also rile liver enzymes. Tight monitoring of liver function makes sense for people taking these drugs, although even "close out-patient supervision and monitoring of liver function" did not prevent acute liver failure (and five deaths) attributed to antiretroviral toxicity at King's College Hospital in London.³³ Still, deaths due to viral hepatitis far outnumber those due to antiretroviral toxicity.

As in the two cases cited by Gallant, uncontrolled hepatitis can lead to cirrhosis and liver cancer. At Stanford University's HIV clinic, Nancy Shulman spotted two cases of hepatocellular carcinoma in people referred to her with well-controlled HIV infection, but with cirrhosis and hepatitis virus coinfection. Both died.

With earlier diagnosis and successful treatment of HCV infection, she e-mailed IAPAC Monthly, both cases "might have been transplantable or operable."

The past year has brought good news on the feasibility -- and success -- of liver transplants in people with HIV infection. One comparison of 24 HIV-infected transplant recipients with age- and race-comparable HIV-uninfected recipients found equivalent survival after 12, 24, and 36 months.³⁴

Alcoholism complicates the management of anyone with HIV infection. And for people who also carry a hepatitis virus, drinking compounds threats to the liver. Hepatitis expert Maurizio Bonacini (California Pacific Medical Center, San Francisco) tells people beginning HCV therapy that the first step is to quit drinking.³⁵ If they can't make that commitment, he believes, they're unlikely to follow other advice.

Because of the shortcomings of anti-HCV therapy, Renslow Sherer (University of Chicago Hospitals) stressed in an e-mail note to IAPAC Monthly, "it is important for patients and doctors to understand that a physician can significantly reduce the risk of chronic liver disease and mortality by persuading and helping the patient to stop all alcohol intake."

No one has the final word on when to treat HCV infection in people with HIV, and the best treatment available today -- pegylated interferon plus ribavirin -- fails all too often. Partly as a result, said some respondents to the survey for this article, HCV-coinfected people often go untreated. But last year two groups -- an ad hoc international panel³⁶ and the British HIV Association³⁷ (BHIVA) proposed thoughtful guidelines for managing HCV in people with HIV. BHIVA also outlined management advice for HIV/HBV-coinfected people.³⁸

Briefly, the international panel makes these points:

1. Screen for HCV antibodies in everyone with HIV infection.

2. People with genotypes 1 and 4 and no or minimal fibrosis may wish to delay therapy until more effective anti-HCV drugs are available.

3. The best candidates for HCV therapy are those with persistently elevated alanine aminotransferase (ALT), a CD4 cell count above 350 cells/mm³, and a viral load below 50,000 copies/ml.

4. People with persistently normal ALT should receive therapy for HCV only if they have fibrosis, or in a clinical trial if they do not have fibrosis.

5. People with a CD4 count below 200 cells/mm³ are generally poor candidates for HCV therapy.

6. People with compensated cirrhosis are candidates for HCV therapy, but those with hepatic decompensation should be considered for liver transplantation.

7. Consider liver transplantation in all people with HCV-related end-stage liver disease and without advanced HIV disease.

8. Substance abuse and severe neuropsychiatric conditions must be addressed before beginning HCV therapy.

Some clinicians try to avert the depressive effects of interferon by coadministering a selective serotonin reuptake inhibitor (SSRI). Johns Hopkins hepatitis expert Stuart Ray starts an SSRI along with interferon in people with a history of alcoholism or a personal or family history of depression.³⁵ In others he starts an SSRI only if a person becomes depressed while taking interferon.

BHIVA suggests treating for HCV coinfection before starting antiretrovirals, when possible, in people with a CD4 count above 200 cells/mm³ because pretreating HCV lowers the risk of liver toxicity when antiretrovirals start.³⁷ For people who have already started antiretrovirals when HCV is diagnosed, BHIVA recommends delaying HCV therapy until the CD4 count climbs over 200 cells/mm³. The British experts advise clinicians that didanosine (ddI) with or without tenofovir (TDF) should be used "with extreme caution" in people taking ribavirin for HCV.

Finally, Johns Hopkins' Gallant reminded IAPAC Monthly, the public health system already has a tool that can prevent infection with hepatitis viruses -- and with HIV as well: needle exchange. Overcoming conservative bias against needle exchange programs in countries such as the United States will take years of concerted effort. But that is no reason not to begin.

Bad Adherence Can Kill

Before people with HIV began taking potent antiretrovirals, the retrovirus killed nearly everyone it infected. In countries that cannot provide antiretrovirals for HIV-infected people, those people are dying today. So it should be no surprise that people who can get antiretrovirals -- but take them haphazardly or not at all -- will die as well.

Anyone who doubts that bad adherence kills might consult two studies in British Columbia, where antiretrovirals are free. The first involved 1,281 antiretroviral-naive adults beginning triple therapy between August 1996 and December 1999.⁹ Defining "intermittent therapy" as filling fewer than 75 percent of prescriptions in the first year of treatment, University of British Columbia researchers used a multivariate model to isolate only two factors that raised the risk of death: Each 100-cell decrement in pretreatment CD4 count raised the risk 1.31 times, and intermittent therapy raised the risk 2.90 times (P<0.001 for both). The researchers ran a subanalysis on people with at least one year of follow-up to eliminate a statistical bias called downward drift -- the chance that less frequent use of antiretrovirals in the first year simply marked more rapid HIV disease progression. This subanalysis confirmed the tie between poor adherence and death.

A more recent and larger study by the same group found that poor adherence outweighed baseline CD4 count in predicting mortality.³⁹ Everyone in the 1,422-person cohort started potent antiretroviral therapy between August 1, 1996 and July 31, 2000. Follow-up continued until March 31, 2002. Again the Vancouver team figured adherence by prescription filling and limited statistical scrutiny to the first year of treatment.

A Kaplan-Meier analysis showed that starting antiretrovirals with 200 or more CD4 cells/mm³ did not prolong survival in the 485 people who had at least 95 percent adherence. But at the 75 percent adherence level, starting treatment even with 350 cells/mm³ or more did not prolong survival in poor adherers. Adjusted relative hazards of death for people with less than 75 percent adherence measured 2.80, 3.05, and 3.21 for people beginning antiretrovirals with 350 cells/mm³ or more, 200 to 349 cells/mm³, and fewer than 200 cells/mm³. The researchers got the same results when they eliminated accidental deaths from the analysis and used the 95 percent adherence cutoff. They believe their findings suggest that bad adherence, rather than how far the baseline CD4 count exceeds 200 cells/mm³, "may be the strongest determinant of patient survival."

Reasons for poor adherence in places with ready access to antiretrovirals could fill a book -- and should. But substance abuse of one kind or another proved the most-cited in the IAPAC Monthly survey. In Miami, Dushyantha Jayaweera (University of Miami) cited crack cocaine as the leading problem; in St. Louis, Judith Aberg (Washington University) named cocaine and alcohol; in San Diego, Mathews sees crystal methamphetamine as "the single biggest barrier to effective HIV care in our clinic," particularly among gay and bisexual men. At Baltimore's Johns Hopkins Hospital, Gallant has an HIV ward "full of people who often die because of their inability to take medications" for a host of reasons -- drug abuse, mental illness, and inadequate housing, to name three.

All HIV clinicians treating poor people in rich countries recognize these often incorrigible problems. From Cincinnati, Carl Fichtenbaum (University of Cincinnati) wrote of the dire need for "comprehensive life-changing programs to alter adherence problems" and "adequate health insurance to allow access to care on an ongoing basis." But sometimes even the best services don't suffice. From the Veterans Administration Medical Center in San Diego, Douglas Richman noted that only six people being treated for HIV died last year -- about 1 percent of the population. They all had "significant emotional and/or substance abuse problems," which prevented them from adhering to antiretroviral regimens despite abundant counseling resources at the facility.

On the other hand, many HIV clinics would consider 1 percent annual mortality a worthy accomplishment. The extra counseling and support that a veterans hospital can offer probably do make regular pill takers out of many people with disorganized lives -- people who might otherwise die. And certainly there are ways to teach adherence to substance abusers. At the IAPAC Sessions 2003, Patricia Kloser (New Jersey Medical School, Newark) explained that "you have to teach that person to be a patient."⁴⁰ She may try to teach adherence by first prescribing a vitamin or, if indicated, PCP prophylaxis, and monitoring the patient closely. Others use initial appointment keeping as a gauge of likely adherence.

Pilot studies of directly observed therapy (DOT) and modified DOT with antiretrovirals show that it can improve adherence and yield virologic benefits in hard-to-treat populations including prisoners,⁴¹ former prisoners,⁴² injecting drug users,^43, 44 and people in methadone maintenance programs.⁴⁵ Antiretroviral DOT has become more feasible with today's once-daily regimens, such as amprenavir (APV)/RTV plus ddI/lamivudine (3TC)⁴⁴ and efavirenz (EFV) with 3TC and abacavir (ABC).⁴⁵

The most common once-daily regimens, Gallant suggested, are:

1. ABC or TDF or ddI, plus

2. 3TC or emtricitabine (FTC), plus

3. EFV or atazanavir (ATV) or ATV/RTV

Such regimens will ease adherence for anyone taking antiretrovirals.

DOT is expensive in developed countries. It requires a dedicated crew of social workers. And how well people continue taking antiretrovirals after DOT stops will only be learned from ongoing studies. But if modified DOT can work in the grinding poverty of rural Haiti,⁴⁶ it should not be written off too readily in poor inner cites and rural regions of wealthy lands.

One can safely bet that all clinicians reading this article became convinced long ago that adherence matters mightily in controlling HIV. Indeed, research shows that clinicians with more HIV experience do better in promoting antiretroviral adherence.⁴⁷ That is only one reason why all should encourage efforts to establish a certificate of qualification that will identify physicians with expertise in HIV infection. But even 20-year veterans of the HIV wars can probably bear to scan the following list of adherence tips, mustered from recent publications and IAPAC Monthly survey responses.

Fateful Impact of a Late Diagnosis

In the IAPAC Monthly survey, late diagnosis of HIV infection tied for second as the most frequently cited cause of death in developed countries today (Table 1). At the Stanford University clinic, Andrew Zolopa sees more than a few people with acute opportunistic infections "who have either been tested and fell through the cracks of the healthcare system or were never tested." From San Diego, Mathews wrote that "we have quite a few undocumented people coming across from Mexico in advanced disease stages who are too sick to survive the initial infections but could have been salvaged if they came in earlier."

Often these "late presenters" come from the same groups that have severe adherence problems, noted Phillip Keiser (Parkland Memorial Hospital, Dallas) -- substance abusers and people with untreated mental illness. These are people "who do not seek regular care," Keiser finds, and "are only seen in the ER when they are sick." Mortal delays in seeking care for HIV seems common throughout the developed world, from Mike Youle's Royal Free Hospital in London, to Jonathan Schapiro's Sheba Medical Center in Tel Aviv, to Pedro Cahn's Fundación Huesped in Buenos Aires. In France, Aquitaine cohort researchers found that 15 percent of cohort members who died in 1998 and 1999 had never taken an antiretroviral.⁴⁹ US researchers attributed excessive HIV-related mortality in Florida prisons to late diagnoses of infection.⁵⁰

Poor public education and little knowledge about HIV explain why some people don't seek care until an advanced opportunistic disease strikes. But a distinct set of HIV-infected people "are aware of their HIV status," wrote Aberg from St. Louis, "but for numerous reasons don't access care and then present with advanced disease." Some of them, she added, still refuse antiretrovirals and even OI prophylaxis, "despite all our educational interventions."

Others avoid the clinic because of poor health coverage -- or embarrassment. Zolopa reported the case of a man who "did not seek out care because of insurance issues" and "ended up with severe cryptococcal disease." He started antiretrovirals and responded well, with undetectable viremia by three months. But he "continued to suffer from complications of his cryptococcal meningitis, leading to blindness and death." In Düsseldorf, Stefan Mauss (Center for HIV and Hepatogastroenterology saw a man with "late-stage anorectal cancer who was shy to present his giant anal condylomata."

The individual clinician can do little for people who doggedly avoid the clinic or disavow care. But sometimes HIV physicians may interact regularly with such diehards and miss the opportunity to encourage testing or treatment. From Chicago, Sherer relayed the case of a man who worked as an HIV care manager in another city. He knew his HIV status for five years but avoided treatment despite daily counseling others on the benefits of therapy. Eventually he returned to Chicago "with seizures and rapidly deteriorated and died with progressive multifocal leukoencephalopathy in eight weeks." Why? "Denial," Sherer suggested, "is a powerful motivator."

From Geneva, Hirschel added that many who deny the gravity of their HIV infection "are influenced by an unrealistically gloomy view of the side effects and complexity of HAART." That sentiment, he proposed, "is fueled by the politically correct line, which is to emphasize side effects and complexities of treatment, so as to keep HIV alive as a serious threat in the minds of the public, and to scare them away from ditching condoms."

People who tenaciously refuse care cannot be helped unless their tenacity can be sapped. Just keeping these people coming back for visits (as Aberg discovered in a case described below) can sometimes pay off. And plenty of people who go undiagnosed because they never suspect they have HIV can surely be stopped well short of death's door. Zolopa believes "programs that integrate testing with care and target high-risk disenfranchised groups could have a very positive impact." Sherer hopes that the new rapid HIV tests "will help both identify infected people and get them immediately into care."

Those are certainly among the goals in the CDC's campaign to expand diagnoses of HIV with a four-part strategy:⁵¹

• Incorporating HIV testing into routine medical care

• Diagnosing HIV infection in nonmedical settings

• Contacting partners of people infected with HIV

• Making HIV testing a regular part of prenatal care and testing all newborns whose mother's HIV status is unknown

Announcing the program, CDC chief Julie Gerberding called it "simply unacceptable that 40,000 people in this country become infected with HIV each year, and it's intolerable that about one fourth of those infected with HIV don't know they're infected and therefore are not receiving appropriate medical care." The CDC figures that one third of those who got a positive HIV test result in 2000 didn't return for treatment.⁵¹ But more than 90 percent with a positive reading on a rapid HIV test came in for a confirmatory assay.⁵²

AIDS Cancers and Non-AIDS Cancers

Most large cohort studies detect a drop in new diagnoses of NHL, and a drop in NHL deaths, since more potent antiretrovirals began bolstering immune systems. But as the proportion of deaths from classic AIDS diagnoses like PCP and toxoplasmosis fell, often-obstinate NHL climbed through the rankings to emerge as a leading killer. Fifteen of 29 physicians responding to the IAPAC Monthly survey rated NHL a leading cause of death in their clinics (Table 1). Non-AIDS cancers ranked just behind NHL in the survey, with 14 clinicians branding these neoplasms as a fatality leader.

A review of the literature on NHL in people with HIV (Table 2) traces several trends since people started taking HAART:

• Sharp drop in incidence (new diagnoses) of NHL after HAART arrived^{17, 54, 56, 59, 63}

• Better response rate to chemotherapy in people responding to HAART^55, 58, 64

• More aggressive anticancer therapy feasible in people taking HAART^{57, 64, 65, 67}

• Longer survival with NHL in people taking HAART^{55, 56, 58, 60, 62, 65, 68}

• Lower NHL incidence and longer survival with HAART linked to higher CD4 counts, lower viral loads, or both^60, 61

During 26,764 person-years of follow-up in the EuroSIDA cohort, the incidence of NHL plunged from 1.99 cases per 100 person-years before September 1995 to 0.30 cases per 100 person-years after March 1999.⁵⁹ HAART appeared to explain the drop. NHL incidence sank from 0.88 cases per 100 person-years during cohort members' first 12 months of HAART to 0.45 cases after more than 24 months (P=0.004). An adjusted Cox model for everyone taking HAART tied the drop in NHL diagnoses to higher CD4 counts and lower viral loads.

In the CASCADE Collaboration study involving 6,941 people with known seroconversion dates from 1979 through 2001, NHL as a first AIDS diagnosis (in 98) dropped by 48 percent among people who picked up HIV in 1997-2001 compared with those infected in 1994-1996 (P<0.05).¹⁷ An analysis of US death certificates found that the proportion of deaths due to NHL rose from 1987 through 1997, surpassing even KS at that point.¹⁹ But the NHL death rate began dwindling after that.

Meanwhile, a raft of studies showed that people with NHL began living longer after they started HAART,^{55, 56, 58, 59, 62, 64, 68} apparently because HAART favored a better response to chemotherapy,^55, 58 enabled people to try more aggressive cancer therapies,^{57, 64, 65, 67} and favored survival by promoting immune recovery.^{59, 60}

But these findings do not hold true in every cohort. Studies of 5,017 Australians with AIDS⁶¹ and of 7,188 HIV-related deaths in the United States²⁴ charted jumps in the proportion of people with NHL as a first diagnosis⁶¹ and in the proportion of NHL deaths²⁴ when comparing HAART years with pre-HAART days (Table 2). Both of these findings may reflect disproportionate drops in first diagnoses of infectious AIDS opportunists and in deaths from opportunistic infections. Even so, such findings probably account for the valid perception that NHL has gained rank as an AIDS-defining disease and as a cause of death since the arrival of HAART. Those are good reasons, Schapiro e-mailed from Tel Aviv, to maintain a "high suspicion for NHL" with "good routine physical exam and labs."

The lingering of NHL as an AIDS diagnosis -- and thus a death threat -- in people with HIV may reflect poor diagnostic skills or this multifarious cancer's complex evolution in people with beleaguered immune systems. Surviving the blitzkrieg attack of infectious opportunists, more people taking antiretrovirals may become eventual targets for lymphoma's war of immune attrition.

A study at London's Chelsea and Westminster Hospital involving 150 HIV-infected people diagnosed with lymphoma since 1986 spotted no drop in lymphoma incidence when comparing pre-HAART years (1988-1995) with the early HAART era (1996-1999).⁶⁶ Probably because new diagnoses of classic opportunists waned after HAART arrived, lymphoma accounted for a significantly bigger percentage of first AIDS diagnoses in the later years (P<0.0001). These researchers found no difference in survival with lymphoma when comparing the two treatment eras, but they did link a lower nadir CD4 count and no prior HAART with development of lymphoma. These NHL predictors, they write, "may translate into a future fall in new cases," as in other cohorts.^{17, 54, 56, 59, 63}

Two important points emerge from studies correlating antiretroviral therapy or anticancer therapy with NHL diagnosis and survival. First, people responding poorly to HAART run a higher risk of NHL and of a quicker death if they do get NHL. At Saint-Louis Hospital in Paris, for example, three quarters of 112 NHL diagnoses in the HAART era happened in people with poorly controlled viremia.⁶⁴ An Italian study of 44 consecutive patients treated simultaneously with HAART and chemotherapy charted complete NHL response rates in 71 percent of HAART responders and 30 percent of nonresponders.⁵⁸ Virologic response to HAART was the only variable these researchers linked to tumor response in a multivariate analysis. Studies like these suggest that people who do poorly with NHL share a trait with people who succumb to other AIDS diseases today: For one reason or another, they cannot or will not stick to their antiretroviral regimen.

Second, HAART may prolong survival with NHL because it promotes a good immune response that makes more aggressive chemotherapy possible. A US study tracked responses in 65 people previously untreated for NHL, giving 40 of them reduced-dose CHOP (cyclophosphamide, doxorubicin, vincristine, and prednisone) and 25 of them full-dose CHOP plus granulocyte colony-stimulating factor.⁵⁷ All patients also took d4T, 3TC, and indinavir (IDV). Complete response rates measured 30 percent in the reduced-dose group and 48 percent in the full-dose group, while side effect rates proved similar in the two groups.

The 112-person study in Paris linked improved survival with more intensive chemotherapy.⁶⁴ And the 44-person Italian study tied higher relative dose intensity of anti-NHL drugs to a reduced risk of death.⁵⁷ Small studies in the United States and Italy found that high-dose chemotherapy and autologous stem cell transplantation are feasible and effective in HIV-infected people with NHL.^65, 67 The Italian researchers propose that "HIV infection should no longer preclude the opportunity of [high-dose therapy] in patients with lymphoma."

The "non-AIDS cancers" may not fit the CDC's definition of opportunistic neoplasms. But for some people with HIV infection today, they pose as grave a threat as KS did in the 1980s. Invasive cervical cancer became such a big threat in women with HIV that the CDC added it to the "AIDS-defining" list. Some think that anorectal cancer deserves the same upgrade.⁷⁰

From Geneva, Hirschel wrote that Hodgkin lymphoma "should be counted among the AIDS-linked cancers" and noted that lung cancer has become more frequent as people with HIV age -- and often continue to smoke. In Madrid, Arribas listed Hodgkin lymphoma, lung cancer, and anorectal cancer -- along with cervical cancer -- as growing threats. Late-stage anorectal cancer killed one of Mauss's patients in Düsseldorf because he was too embarrassed to seek care for large anal condylomata.

Cohort studies and other surveys that count deaths from non-AIDS cancers suggest that 10 to 20 percent of deaths in the HAART years may be traced to these malignancies, depending on how deaths are counted. Analysis of 107 deaths in the French Aquitaine cohort in 1998 and 1999 attributed 12 of them (11 percent) to non-AIDS cancers.⁸ Non-AIDS malignancies accounted for 13 percent of non-AIDS-defining causes of death from 1994 through 2000 in the US Women's Interagency HIV Study (WIHS) cohort.²³ Analysis of 863 deaths in France among HAART responders with a CD4 count above 200 cells/mm³ and a viral load under 500 copies/mL blamed non-AIDS cancers (usually lung or anorectal cancer) for 19 percent.⁷¹

The proportion of deaths due to non-AIDS cancers grew significantly in a San Francisco survey of 5,234 HIV-related deaths from 1994 (6.4 percent) through 1998 (10.9 percent, P<0.01).⁷² The main non-AIDS cancers were lung cancer, Hodgkin disease, liver cancer, and anal cancer. The US death certificate analysis of people with HIV found a stable proportion of deaths from cancers of the lung, bronchus, and trachea through 1995, an increase in 1996 that peaked in 1998, and a fall in 1999.¹⁹ Because some of these studies count hepatocellular carcinoma among non-AIDS cancers, they include deaths that can be traced back to hepatitis virus coinfection.

How Big Is the Threat From Heart Disease?

The risk of cardiovascular disease in people with HIV infection grabbed the attention of many protease inhibitor prescribers and takers when it became clear that most PIs -- especially with a lift from RTV -- can send lipids skyward. (Atazanavir is an exception.) Big cohort studies linking a growing risk of myocardial infarction with more years taking PIs⁷⁶ or combination antiretroviral therapy (with or without PIs)⁷⁷ stoked those concerns.

But heart disease is no new affliction in people with HIV infection. By 1999 heart disease had become the fifth leading cause of death among people with HIV, according to a US death certificate survey (trailing "unspecified pneumonia," septicemia, liver disease, and kidney disease).¹⁹ Among non-AIDS deaths from 1994 through 2000 in the US Women's Interagency HIV Study, cardiac disease tied for fourth place (along with accident, suicide, or homicide) by accounting for 11 percent of deaths.²³ In the 2000 French HIV mortality study, cardiovascular disease led to 16 percent of deaths in people responding well to HAART.⁷¹

And the rate of heart-related deaths among people with HIV is on the rise. A survey of six US cities found a (nonsignificant) 1.9 times higher risk of death from ischemic heart disease in 1996-2000 than in 1992-1995.²⁴ A San Francisco study figured that the proportion of deaths from coronary artery disease in people with HIV rose from 0.8 percent in 1994 and 0.5 percent in 1995 to 1.2 percent in 1996, 1.3 percent in 1997, and 1.8 percent in 1998 (P<0.01).⁷² A EuroSIDA study stretching from 1994 to 2000-2001 ranked myocardial infarction as the second leading non-AIDS, nonsuicide cause of death (after hepatitis) since January 2000.⁴

A study looking at serious or life-threatening complications among 2,947 people enrolled in five clinical trials from 1996 through 2001 rated cardiovascular complications just behind liver problems and neutropenia/anemia.⁶ But the risk of death from heart disease was higher than with any other complication (hazard ratio 7.08, P=0.0001), just higher than the death risk from AIDS (hazard ratio 6.95, P=0.00001).

Heart disease may be making bigger blips on the death risk radar screen, yet only eight of the 29 clinicians IAPAC Monthly surveyed named it as a mortality leader in their clinic, putting cardio complications sixth on their list of death threats (Table 1). That level of concern reflects the low incidence of myocardial infarction in the multicohort DAD study (3.5 per 1,000 person-years)⁷⁷ and the French hospital study (1.6 per 1,000 person-years for men taking PIs for 18 to 29 months, and 3.4 per 1,000 person-years for those taking PIs longer, compared with an expected rate of 1.1 per 1,000 person-years in the French male population).⁷⁶ Of the 126 infarctions counted in DAD, only 36 (29 percent) proved fatal, and only 6 percent of deaths in the cohort resulted from infarction. The French team did not report how many of the 60 heart attacks in their analysis were fatal.

Clinicians at Boston's Brigham and Women's Hospital have seen only a few cardiovascular deaths in people with HIV, Kuritzkes told IAPAC Monthly. "They've generally been in people who have other underlying risk factors -- diabetes, hypertension, and kidney disease," Kuritzkes says. "And it's generally in people who are of an age where you would expect to begin seeing cardiovascular mortality, which makes it very hard to sort out if it's an increased risk due to antiretroviral therapy or HIV infection."

A recent survey of a US managed care database backs Kuritzkes' clinical impression. Analyzing 174 hospital admissions of 18- to 90-year-olds with HIV between January 1 and June 30, 2000, the University of Cincinnati's Fichtenbaum found that 42.5 percent admitted for heart problems had hypertension and 22.5 percent had diabetes.⁷⁸ Four of five people with heart disease were men, and their age at admission averaged 55.6 years.

Although research ties both hypertension and diabetes to PI therapy, a third big contributor to heart disease has nothing to do with antiretrovirals -- smoking. Among 535 people with HIV who died in France from January to June 2000, 51 percent smoked.¹⁴ Cardiovascular disease came in third among non-AIDS causes of death in this survey. Current or former smoking, along with older age, previous heart disease, and male gender, independently predicted myocardial infarction in DAD.⁷⁷

If heart disease has yet to vault into the first rank of mortality concerns among HIV clinicians, the growing myocardial infarction risk over time in the French study⁷⁶ and DAD,⁷⁷ along with an aging HIV population in Western countries, raises the odds that cardiovascular complications will climb up the concern list in this decade. In Fichtenbaum's managed care database population,⁷⁸ heart disease already sends more HIV-infected people to the hospital than liver disease, kidney disease, or opportunistic infection. In fact, significantly more people checked into the hospital with heart problems than with liver complications (P<0.05). Only nonopportunistic infections beat cardiovascular disease in this hospital admission contest.

At London's Chelsea and Westminster Hospital, Moyle figures that heart disease causes one in 20 deaths among people with HIV. He tries to stick to National Cholesterol Education Program (NCEP) guidelines⁷⁹ in managing high lipids, a task that "may involve earlier switching approaches, most notably away from Kaletra." Fehmida Visnegarwala (Baylor College of Medicine) made a similar point from her clinic in Houston, writing that "cardiovascular causes [of death] can be prevented with careful patient education at the start of HAART therapy and avoiding drugs such as the boosted protease inhibitors as a first-line option among those with an elevated cardiovascular risk." At Washington University in St. Louis, Aberg worries about ischemia, cardiomyopathy, and arrhythmias in the cocaine abusers she treats.

Dabbling in HIV Medicine, Dabbling in Death

Seven of the 29 clinicians who answered the IAPAC Monthly survey listed "care by clinicians with limited HIV experience" as a leading contributor to deaths in their clinic. Although most survey respondents did not see meager experience as an immediate cause of death, clearly it can contribute to any of the other causes listed in Table 1. Several studies back that impression statistically.

Before the arrival of potent antiretrovirals, a landmark study of 403 men who got an AIDS diagnosis from 1984 through mid-1994 found that those treated by physicians with more HIV experience lived longer.¹⁰ After controlling for severity of illness and year of diagnosis, researchers at the University of Washington in Seattle determined that men cared for by clinicians with the most HIV experience had a 31 percent lower risk of death than men seen by clinicians with the least experience (P<0.02).

Putting more powerful drugs into inexperienced clinicians' hands didn't solve this problem. Indeed, it may have made things worse by fueling the evolution of drug-resistant virus. In Germany, Mauss told IAPAC Monthly he has seen "a number of patients showing up with full-blown AIDS and no antiretroviral options left due to long-term treatment with failing regimens by ignorant or inexperienced physicians."

The daunting task of interpreting resistance tests contributes to this poor care, according to Kuritzkes. "I think we're beginning to plateau with resistance testing," he explained, "in just how well we can reach and educate providers who are very busy doing general medicine as well as some HIV care." These physicians typically "are unable to get to HIV meetings, don't have time to read and develop a deep understanding of resistance testing, and they write a large number of prescriptions."

A survey of clinicians attending an HIV symposium in New York City disclosed a dismal grasp of basic resistance principles,⁸¹ even though one might assume that physicians paying to hear HIV talks would have more than a glimmer of resistance savvy. A similar survey at another symposium two years later showed only modest improvement.⁸² Although 57 percent of respondents claimed to have "moderate to high expertise" in treating HIV infection and 22 percent rated themselves HIV "experts," only 43 percent could name a mutation that caused resistance to five of seven antiretroviral groups, and 27 percent failed to make a correct match in every category.

So it's no surprise that a more recent study of physicians' HIV experience by the University of Washington group again linked less experience to worse survival odds.¹² Among HIV-infected people cared for at a large health maintenance organization from 1990 through 1999, those who saw physicians with the least HIV experience got the lowest level of outpatient pharmacy and laboratory services (P<0.001) and were half as likely to visit a specialist (P<0.05) as people seeing a physician with moderate or more HIV experience. And the people who got low-level services were 15 times more likely to die during the course of the study (P=0.02). In British Columbia, a province with universal access to antiretrovirals, a study of 1,282 antiretroviral-naive people starting triple therapy between August 1996 and December 1999 found that every additional 100 HIV patients in a clinician's practice lowered the risk of death by almost 25 percent (P=0.037).⁹

In some parts of the United States, particularly rural areas, people with HIV have to travel far to find a top HIV clinic. Many don't make the trip. In Colorado, Kuritzkes told IAPAC Monthly, HIV savvy ran thin outside a few urban centers. As a stopgap, every week one of his colleagues at the University of Colorado "would fly out to Grand Junction to hold HIV rounds with people who were holding down the fort for the rest of the week." Even when people live near seasoned HIV practitioners, he added, they can remain "stubbornly loyal to physicians who are well meaning but not well qualified."

Treatment by physicians with scant HIV skills is a tough problem to solve, but there are ways. One, suggested Gallant from Johns Hopkins, is to set a consistent policy. "Lack of care by experts needs to be addressed nationally," he wrote in responding to the IAPAC Monthly survey. "On the one hand the government says (in the DHHS guidelines⁴⁸) that HIV care should be managed by experts. On the other hand we spend lots of money (through [the Health Resources Service Administration]) funding programs to teach generalists how to dabble in HIV. We need to get a consistent message out that HIV care should be managed (or at least co-managed) by experts." Putting pressure on third-party payers and managed care organizations, which often resist expert consultations, and educating people with HIV to insist on expert care would also pay dividends, Gallant added.

Several physicians' groups in the United States -- including the HIV Medicine Association (HIVMA) and International Association of Physicians in AIDS Care (IAPAC) -- are trying to set standards that earmark experts in HIV care. The HIVMA and IAPAC are working together with the American Board of Internal Medicine and other boards to develop a Certificate of Added Qualification that will identify clinicians expert in HIV medicine. Kuritzkes, now HIVMA's vice chairman, said his group also helps develop "training pathways for people who are not doing infectious disease fellowships and want to do HIV care."

German HIV mavens devised a way to address shortcomings in physicians' drug resistance knowledge. The Radata Project now embraces 59 HIV centers and 15 laboratories that collaborate to collate clinical data, resistance test and therapeutic drug monitoring results, and patient self-reports in an Internet database.^83, 84 Physicians treating people with HIV and considering a regimen switch may contact Radata to get advice from two independent experts. That patient's data must be submitted to the database. At last year's 9th European AIDS Conference, Radata members reported that clinicians have sought switch advice 663 times for 193 people so far.⁸³

In reviewing a draft of this article, Sherer stressed the need "for ongoing mentoring and consultation as an integral part of an HIV care system. HIV medicine has become so much more complex that ongoing training updates and access to real-time consultations are important adjuncts to care for any HIV clinician, including the so-called experts."

... And All Those Other Reasons

So far this review has considered the seven leading causes of death in HIV-infected people with access to antiretrovirals, as judged by a panel of 29 HIV clinicians. But other mortal threats abound. A literature review suggested 12 other prime causes of death in people with HIV, and at least two survey respondents voted for each of them (Table 1). In Providence, Rhode Island, Charles Carpenter (Brown University) is "most concerned about the high rates of renal failure, mostly in African Americans." In Chicago, Sherer has seen suicides and deaths resulting from domestic violence, especially involving women and gay teens.

Some physicians who answered the IAPAC Monthly survey cited emerging causes of death that have received little or no attention in the literature. "We've seen several cases of fatal pulmonary embolism in the setting of what appears to be a hypercoagulable state," Zolopa wrote from Stanford. "Whether this is drug related or disease related is unclear." In Los Angeles, Peter Ruane (Tower Infectious Diseases Medical Associates) has come across a growing number of fatal atypical cytomegalovirus infections and invasive aspergillosis.

Meanwhile, the grisly opportunistic infections that quickly killed thousands in the years before effective prophylaxis and potent antiretrovirals have not gone away. A chart review comparing causes of death in 1995 and 1999-2000 at Parkland Memorial Hospital in Dallas counted 21 deaths from PCP in the first period (19 percent) and 15 in the second (17 percent).⁷ In 1995, 77 people (69 percent) had taken PCP prophylaxis, compared with 44 (50 percent) in 1999-2000 (P<0.01). The authors note that poor antiretroviral adherence appeared to correlate with poor prophylaxis adherence in the needy minority population that the hospital serves.

A few clinicians reminded IAPAC Monthly that antiretrovirals themselves can be the proximate cause of death. From Madrid, Arribas wrote that "we had a death caused by an interaction between lopinavir/ritonavir [LPV/RTV] and an ergot derivative." From Chicago, Sherer related the case of a man in another city who had been taking efavirenz for one week, became psychotic, and during a visit to his clinician threw himself from the eighth-floor office window.

Lactic acidosis was the most commonly cited antiretroviral toxicity in the IAPAC Monthly survey. Despite the growing attention this fatal complication has earned in the past several years, the symptoms sometimes still go unaddressed until it's too late. From Houston, Visnegarwala reported the case of a 37-year-old woman taking nelfinavir (NFV), d4T, and 3TC. Nonspecific abdominal pain led to two months of treatment for "fibroid disease." When first seen at Visnegarwala's clinic, the woman suffered from weight loss, shortness of breath, and a lactate level of 8.0 mmol/L. Despite intensive care including dialysis, intravenous carnitine, and multivitamin infusion, she died.

Although treatment guidelines typically do not recommend routine lactate monitoring, some HIV clinicians have taken up the practice, including Moyle in London.

He e-mailed IAPAC Monthly that "serial monitoring may provide early pick-up of some events." Yet no one doubts that, properly used, the potent antiretrovirals now available have saved far more lives than they threaten. Last year saw the licensing of five new antiretrovirals in the United States, including agents that are easier to take (fosamprenavir, NFV 625 mg, ATV), sometimes less toxic (ATV, fosamprenavir, FTC), and target a different step in the viral life cycle (enfuvirtide [ENF]).

Even in people with advanced disease, combinations of some new and some old drugs can still turn the tables on HIV. But first the drugs have to be prescribed, then they have to be taken. EuroSIDA researchers traced a steady correlation between treatment with more drugs and a lower risk of progression or death in 1,106 people with CD4 counts under 50 cells/mm³.⁸⁵ Compared with people taking five drugs, the risk grew progressively greater in people taking four, three, two, and no drugs. Earlier research by EuroSIDA workers showed that continuing a PI regimen, even in people with fewer than 50 cells/mm³, cut the risk of progression 43 percent compared with not taking a PI when the CD4 count falls that low.⁸⁶

More recent work by the 13-cohort PLATO collaboration showed that CD4 counts continued to climb in people with triple-class failure as long as their viral load stayed below 10,000 copies/mL or 2 logs below the pretreatment setpoint.⁸⁷ Even in people with no apparent virologic suppression from continued therapy, the CD4 count dropped more slowly than it did in people who abandoned treatment. Further research by the PLATO group found a 3-fold higher risk of death in people with triple-class failure if their CD4 count fell below 200 cells/mm³ and a 16-fold higher risk if it fell below 50 cells/mm³.⁸⁸ Not taking antiretrovirals independently raised the risk of death 2.85 times.

On the other hand, a 1996-1999 survey by University of British Columbia researchers correlated intermittent use of triple therapy (judged by pharmacy records) in the first year of treatment with a 2.9 times higher risk of death (P<0.001).⁹

What can new regimens do for people with advanced infection and a record of treatment failures? From Toronto, Mona Loutfy (University of Toronto) sketched the case of a person with multidrug-resistant virus and a CD4 count below 50 cells/mm³ who started LPV/RTV, APV, ddI, 3TC, TDF, and T-20. The T-cell count climbed above 400 cells/mm³ and the viral load sank below 50 copies/mL. But the patient died of a cocaine overdose.

In London, Youle saw a 25-year-old who never could stomach antiretrovirals for more than six weeks, suffering from nausea, diarrhea, and mental health problems. His T-cell tally stood at 22 cells/mm³ and his viral load at 300,000 copies/mL. Eight weeks after starting ATV/RTV plus TDF and ddI, his viral load has dropped 2 logs, he feels no side effects, "and seems to be keen to continue." Cases like these make an obvious point: Don't give up on people with advanced disease -- and don't let them give up on themselves. A page from Aberg's casebook offers one more example:

A 25-year-old woman infected since she was 17 years old had a CD4 count of 99 cells/mm³ and difficulty adhering to any regimen. Cervical dysplasia, oral candidiasis, and bacterial pneumonia complicated her course. In March 2000, her clinicians diagnosed toxoplasmosis and central nervous system lymphoma. She took drugs for the toxo for two months but refused radiation therapy and all other medications.

Finally, she gave up and started planning her funeral. But she kept her medical appointments. In September 2000, she met another patient in the waiting room of Aberg's clinic, fell in love, and decided to give antiretrovirals another try. Her brain lesions resolved, her viral load has been out of sight since December 2000, her CD4 count now tops 800 cells/mm³, and she delivered a healthy baby boy in August 2002.

Mark Mascolini is the IAPAC Monthly's Writer-at-Large. He may be reached at mailmark@ptd.net.

Acknowledgments

Thanks to Renslow Sherer (University of Chicago Hospitals) for reviewing this article, to Joel Gallant (Johns Hopkins University, Baltimore) and Bernard Hirschel (University Hospital, Geneva) for extensive comments, and to Daniel Kuritzkes (Brigham and Women's Hospital, Boston) and Alan Landay (Rush Presbyterian-St. Luke's Medical Center, Chicago) for interviews. Fabrice Bonnet (Saint-André Hospital, Bordeaux, France), Mardge Cohen (Cook County Hospital, Chicago), Joel Palefsky (University of California, San Francisco), and Richard Selik (US Centers for Disease Control and Prevention, Atlanta) provided reprints.

The following clinicians responded to the survey:

Judith Aberg (Washington University, St. Louis, Missouri), José Arribas (Hospital La Paz, Madrid), Pedro Cahn (Fundación Huesped, Buenos Aires), Charles Carpenter (Brown University, Providence, Rhode Island), Bonaventura Clotet (Fundacio IRSI Caixa, Barcelona, Spain), Carl Fichtenbaum (University of Cincinnati), Joel Gallant (Johns Hopkins University, Baltimore, Maryland), Bernard Hirschel (University Hospital Geneva), Dushyantha Jayaweera (University of Miami), Philip Keiser (University of Texas Southwestern Medical School, Dallas), Daniel Kuritzkes (Brigham and Women's Hospital, Boston), Mona Loutfy (University of Toronto), Franco Maggiolo (Ospedali Riuniti, Bergamo, Italy), Josep Mallolas (University of Barcelona, Spain), Roberto Manfredi (University of Bologna "Alma Mater Studiorum"), W. Christopher Mathews (University of California, San Diego), Stefan Mauss (Center for HIV and Hepatogastroenterology, Düsseldorf, Germany), Graeme Moyle (Chelsea and Westminster Hospital, London), Ronald Reisler (University of Maryland, Baltimore), Peter Reiss (Academic Medical Center, Amsterdam), Douglas Richman (University of California, San Diego), William Rodriguez (Brigham and Women's Hospital, Boston), Peter Ruane (Tower Infectious Diseases Medical Associates, Los Angeles), Jonathan Schapiro (Sheba Medical Center, Tel Aviv), Renslow Sherer (University of Chicago Hospitals), Nancy Shulman (Stanford University, Stanford, California), Fehmida Visnegarwala (Baylor College of Medicine, Houston), Mike Youle (Royal Free Hospital, London), Andrew Zolopa (Stanford University, Stanford, California).

References and Notes

1. Kaiser Daily HIV/AIDS Report. Three people with HIV/AIDS die while on West Virginia ADAP waiting list. August 29, 2003. www.thebody.com/kaiser/2003/aug29_03/west_va_adap.html

2. CDC Division of HIV/AIDS Prevention. U.S. HIV and AIDS cases reported through June 2000. Midyear edition Vol. 12, No. 1. Table 27. Estimated deaths of persons with AIDS, by race/ethnicity and year of death, 1993 through 1999, United States. www.cdc.gov/hiv/stats/hasr1201/table26.htm#fig26

3. CDC Division of HIV/AIDS Prevention. Cases of HIV infection and AIDS in the United States, 2002. HIV/AIDS Surveillance Report, Vol. 14. Table 7. Estimated numbers of deaths of persons with AIDS, by year of death and selected characteristics, 1998-2002 -- United States. www.cdc.gov/hiv/stats/hasr1402/table7.htm

4. Mocroft A, Brettle R, Kirk O, et al. Changes in the cause of death among HIV positive subjects across Europe: results from the EuroSIDA study. AIDS 2002;16:1663-1671.

5. Mathews W, Barber E. Rising mortality rates among recent entrants to care for HIV infection. 10th Conference on Retroviruses and Opportunistic Infections. February 10-14, 2003. Boston. Abstract 911.

6. Reisler RB, Han C, Burman WJ, et al. Grade 4 events are as important as AIDS events in the era of HAART. JAIDS 2003;34:379-386.

7. Jain MK, Skiest DJ, Cloud JW, et al. Changes in mortality related to human immunodeficiency virus infection: comparative analysis of inpatient deaths in 1995 and in 1999-2000. Clin Infect Dis 2003;36:1030-1038.

8. Bonnet F, Morlat P, Chêne G, et al. Causes of death among HIV-infected patients in the era of highly active antiretroviral therapy, Bordeaux, France, 1998-1999. HIV Medicine 2002;3:195-199.

9. Hogg RS, Heath K, Bangsberg D, et al. Intermittent use of triple-combination therapy is predictive of mortality at baseline and after 1 year of follow-up. AIDS 2002;16:1051-1058.

10. Kitahata MM, Koepsell TD, Deyo RA, et al. Physicians' experience with the acquired immunodeficiency syndrome as a factor in patients' survival. N Engl J Med 1996;334:701-706.

11. Laine C, Markson LE, McKee LJ, et al. The relationship of clinic experience with advanced HIV and survival of women with AIDS. AIDS 1998;12:417-424.

12. Kitahata MM, Van Rompaey SE, Dillingham PW, et al. Primary care delivery is associated with greater physician experience and improved survival among persons with AIDS. J Gen Intern Med 2003;18:95-103.

13. Valdez H, Chowdhry TK, Assad R, et al. Changing spectrum of mortality due to human immunodeficiency virus: analysis of 260 deaths during 1995-1999. Clin Infect Dis 2001;32:1487-1493.

14. Green L, Cohen MH, Pulvirenti J, Sherer R. Changing trends in causes of death by gender in Chicago since HAART, 1996-1999. XIV International AIDS Conference. July 7-12, 2002. Barcelona. Abstract MoPeC3337.

15. Lewden C, Héripret L, Bonnet F, et al. Causes of death in HIV-infected adults in the era of highly active antiretroviral therapy (HAART): the French survey "Mortalité 2000." 9th Conference on Retroviruses and Opportunistic Infections. February 24-28, 2002. Seattle. Abstract 753.

16. Folch E, Hernández I, Vetter J, et al. AIDS-related mortality: improved survival with HAART but causes of death unchanged. 9th Conference on Retroviruses and Opportunistic Infections. February 24-28, 2002. Seattle. Abstract 756.

17. CASCADE Collaboration. Changes over calendar time in the risk of specific first AIDS-defining events following HIV seroconversion, adjusting for competing risks. Int J Epidemiol 2002;31:951-958.

18. The "other" AIDS-defining illnesses in the CASCADE study were invasive cervical carcinoma, coccidioidomycosis, isosporiasis, and salmonella septicemia.

19. Selik RM, Byers RH Jr, Dworkin MS, et al. Trends in diseases reported on U.S. death certificates that mentioned HIV infection, 1987-1999. JAIDS 2002;29:378-387.

20. Nuesch R, Geigy N, Schaedler E, Battegay M. Effect of highly active antiretroviral therapy on hospitalization characteristics of HIV-infected patients. Eur J Clin Microbiol Infect Dis 2002;21:684-687.

21. Jaggy C, von Overbeck J, Ledergerber G, et al. Mortality in the Swiss HIV Cohort Study (SHCS) and the Swiss general population. Lancet 2003;362:877-878.

22. Rodríguez B, Valdez H, Salata R, et al. Changes in the causes of death in a cohort of HIV-infected individuals: analysis of the last 6 years. 9th Conference on Retroviruses and Opportunistic Infections. February 24-28, 2002. Seattle. Abstract 755.

23. Cohen MH, French AL, Benning L, et al. Causes of death among women with human immunodeficiency virus infection in the era of combination antiretroviral therapy. Am J Med 2002;113:91-98.

24. Wolfe MI, Hanson DL, Selik R, Swerdlow DL. Deaths from non-AIDS-related diseases have increased as a proportion of deaths of HIV-infected persons since the advent of HAART. 9th Conference on Retroviruses and Opportunistic Infections. February 24-28, 2002. Seattle. Abstract 14.

25. McGovern B, Bica I, Snydman D, et al. Mortality due to hepatitis C virus-related cirrhosis in patients infected with HIV type 1: a role for alcohol. Reply. Clin Infect Dis 2001;33:1796-1797.

26. Bica I, McGovern B, Dhar R, et al. Increasing mortality due to end-stage liver disease in patients with human immunodeficiency virus infection. Clin Infect Dis 2001;32:492-497.

27. Pérez-Olmeda M, Soriano V, Asensi V, et al. Treatment of chronic hepatitis C in HIV-infected patients with interferon alpha-2b plus ribavirin. AIDS Res Hum Retroviruses 2003;19:1083-1089.

28. US Public Health Service and Infectious Diseases Society of America. 2001 USPHS/IDSA guidelines for the prevention of opportunistic infections in persons infected with human immunodeficiency virus. November 28, 2001. www.aidsinfo.nih.gov/guidelines/default_db2.asp?id=69

29. At IAPAC Sessions 2002, Barbara McGovern (New England Medical Center, Boston) suggested delaying HBV vaccination in people with a CD4 count below 200 cells/mm³ until antiretrovirals push the count above that mark. Stuart Ray (Johns Hopkins University, Baltimore) agreed but urged colleagues to set a vaccination deadline six or 12 months after starting antiretrovirals regardless of CD4 count to prevent procrastination from robbing these people of whatever benefit they may derive from the vaccine. The British HIV Association (BHIVA) recommends revaccinating HCV-infected people who had an earlier HBV vaccine failure if antiretrovirals raise their CD4 count.³⁷

30. Sulkowski MS, Thomas DL, Chaisson RE, et al. Hepatotoxicity associated with antiretroviral therapy in adults infected with human immunodeficiency virus and the role of hepatitis C or B virus infection. JAMA 2000;283:74-80.

31. Qurishi N, Kreuzberg C, Lüchters G, et al. Effect of antiretroviral therapy on liver-related mortality in patients with HIV and hepatitis C virus coinfection. Lancet 2003;362:1708-1713.

32. Gisolf EH, Dreezen C, Danner SA, et al. Risk factors for hepatotoxicity in HIV-1-infected patients receiving ritonavir and saquinavir with or without stavudine. Clin Infect Dis 2000;31:1234-1239.

33. Clark SJ, Creighton S, Portmann B, et al. Acute liver failure associated with antiretroviral treatment for HIV: a report of six cases. J Hepatol 2002;36:295-301.

34. Ragni MV, Belle SH, Im K, et al. Survival of human immunodeficiency virus-infected liver transplant recipients. J Infect Dis 2003;188:1412-1420.

35. Mascolini M. Clinicians debate antiretroviral tactics and toxicities, resistance testing, and hepatitis virus coinfection. IAPAC Sessions 2002. IAPAC Monthly 2002;8:190-201. www.thebody.com/iapac/jul02/clinicians_debate.html

36. Soriano V, Puoti M, Sulkowski M, et al. Care of patients with hepatitis C and HIV co-infection. AIDS 2004;18:1-12.

37. BHIVA Coinfection Guideline Committee on behalf of the British HIV Association. British HIV Association (BHIVA) guidelines for treatment and management of HIV and hepatitis C coinfection. October 2003. www.bhiva.org/guidelines/2003/HCV/index.html

38. BHIVA Coinfection Guideline Committee on behalf of the British HIV Association. BHIVA hepatitis coinfection guidelines. October 2003. www.bhiva.org/guidelines/2003/HBV/index.html

39. Wood E, Hogg RS, Yip B, et al. Effect of medication adherence on survival of HIV-infected adults who start highly active antiretroviral therapy when the CD4+ cell count is 0.200 to 0.350 x 109 cells/L. Ann Intern Med 2003;139:810-816.

40. Mascolini M. Keeping an eye on HIV treatment: from substance abuse to side effects. IAPAC Sessions 2003. IAPAC Monthly 2003;9:146-165. www.thebody.com/iapac/jul03/treatment.html

41. Fischl M, Castro J, Monroig R, et al. Impact of directly observed therapy on long-term outcomes in HIV clinical trials. 8th Conference on Retroviruses and Opportunistic Infections. February 4-8, 2001. Chicago. Abstract 528.

42. Stenzel MS, McKenzie M, Mitty JA, Flanigan TP. Enhancing adherence to HAART: a pilot program of modified directly observed therapy. AIDS Read 2001;11:317-319, 324-328.

43. Altice FL, Mezger J, Jodges J, et al. Directly observed therapy for HIV+ drug users. 2nd IAS Conference on HIV Pathogenesis and Treatment. July 13-16, 2003. Paris. Abstract 40.

44. Garg V, Brill M, Rodriguez A, et al. Clinical pharmacology of amprenavir in highly challenging, naive patients following a once daily HAART regimen under directly observed therapy (DOT). 3rd International Workshop on Clinical Pharmacology and HIV Therapy. April 11-13, 2003. Washington, DC. Abstract 2.5.

45. Pulik P, Higersberger J, Bednarska A, et al. Preliminary results of DOT-POL study -- daily observed therapy with lamivudine, abacavir, efavirenz (300/600/600 mg) QD in HIV-1 infected patients participating in methadone maintenance programme. 9th European AIDS Conference. October 25-29, 2003. Warsaw. Abstract 7.3/18.

46. Farmer P, Leandre F, Mukherjee JS, et al. Community-based approaches to HIV treatment in resource-poor settings. Lancet 2001;358:404-409.

47. Delgado J, Heath KV, Yip B, et al. Highly active antiretroviral therapy: physician experience and enhanced adherence to prescription refill. Antivir Ther 2003;8:471-478.

48. Panel on Clinical Practices for Treatment of HIV Infection convened by the Department of Health and Human Services. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. November 10, 2003. www.aidsinfo.nih.gov/guidelines/default_db2.asp?id=50

49. Bonnet F, Morlat P, Chêne G, et al. Causes of death among HIV-infected patients in the era of highly active antiretroviral therapy, Bordeaux, France, 1998-1999. HIV Medicine 2002;3:195-199.

50. Amankwaa AA, Bavon AL, Amankwaa LC. Gaps between HIV/AIDS policies and treatment in correctional facilities. J Health Hum Serv Adm 2001;24:171-198.

51. Centers for Disease Control and Prevention. Advancing HIV prevention: new strategies for a changing epidemic -- United States, 2003. MMWR 2003;52(15):329-332. www.cdc.gov/MMWR/PDF/wk/mm5215.pdf

52. Barclay L. CDC initiative for HIV prevention: a newsmaker interview with Harold W. Jaffe, MD. Medscape HIV/AIDS. www.medscape.com/viewarticle/452858

53. Ledergerber B, Talenti A, Egger M. Risk of HIV related Kaposi's sarcoma and non-Hodgkin's lymphoma with potent antiretroviral therapy: prospective cohort study. BMJ 1999;319:23-24.

54. International Collaboration on HIV and Cancer. Highly active antiretroviral therapy and incidence of cancer in human immunodeficiency virus-infected adults. J Natl Cancer Inst 2000;92:1823-1830.

55. Navarro JT, Ribera JM, Oriol A, et al. Influence of highly active anti-retroviral therapy on response to treatment and survival in patients with acquired immunodeficiency syndrome-related non-Hodgkin's lymphoma treated with cyclophosphamide, hydroxydoxorubicin, vincristine and prednisone. Br J Haematol 2001;112:909-915.

56. Chow KU, Mitrou PS, Geduldig K, et al. Changing incidence and survival in patients with AIDS-related non-Hodgkin's lymphomas in the era of highly active antiretroviral therapy (HAART). Leuk Lymphoma 2001;41:105-116.

57. Ratner L, Lee J, Tang S, et al. Chemotherapy for human immunodeficiency virus-associated non-Hodgkin's lymphoma in combination with highly active antiretroviral therapy. J Clin Oncol 2001;19:2171-2178.

58. Antinori A, Cingolani A, Alba L, et al. Better response to chemotherapy and prolonged survival in AIDS-related lymphomas responding to highly active antiretroviral therapy. AIDS 2001;15:1483-1491.

59. Kirk O, Pedersen C, Cozzi-Lepri A, et al. Non-Hodgkin lymphoma in HIV-infected patients in the era of highly active antiretroviral therapy. Blood 2001;98:3406-3412.

60. Baiocchi OC, Colleoni GW, Navajas EV, et al. Impact of highly active antiretroviral therapy in the treatment of HIV-infected patients with systemic non-Hodgkin's lymphoma. Acta Oncol 2002;41:192-196.

61. Dore GJ, Li Y, McDonald A, et al. Impact of highly active antiretroviral therapy on individual AIDS-defining illness incidence and survival in Australia. JAIDS 2002;29:388-395.

62. Tam HK, Zhang ZF, Jacobson LP, et al. Effect of highly active antiretroviral therapy on survival among HIV-infected men with Kaposi sarcoma or non-Hodgkin lymphoma. Int J Cancer 2002;98:916-922.

63. Eltom MA, Jemal A, Mbulaiteye SM, et al. Trends in Kaposi's sarcoma and non-Hodgkin's lymphoma incidence in the United States from 1973 through 1998. J Natl Cancer Inst 2002;94:1204-1210.

64. Gerard L, Galicier L, Maillard A, et al. Systemic non-Hodgkin lymphoma in HIV-infected patients with effective suppression of HIV replication: persistent occurrence but improved survival. JAIDS 2002;30:478-484.

65. Molina A, Krishnan AY, Nademanee A, et al. High dose therapy and autologous stem cell transplantation for human immunodeficiency virus-associated non-Hodgkin lymphoma in the era of highly active antiretroviral therapy. Cancer 2000;89:680-689.

66. Matthews GV, Bower M, Mandalia S, et al. Changes in acquired immunodeficiency syndrome-related lymphoma since the introduction of highly active antiretroviral therapy. Blood 2000;96:2730-2734.

67. Re A, Cattaneo C, Michieli M, et al. High-dose therapy and autologous peripheral-blood stem-cell transplantation as salvage treatment for HIV-associated lymphoma in patients receiving highly active antiretroviral therapy. J Clin Oncol 2003;21:4423-4427.

68. Vaccher E, Spina M, Talamini R, et al. Improvement of systemic human immunodeficiency virus-related non-Hodgkin lymphoma outcome in the era of highly active antiretroviral therapy. Clin Infect Dis 2003;37:1556-1564.

69. Spina M, Carbone A, Vaccher E, et al. Outcome in patients with non-Hodgkin lymphoma and with or without human immunodeficiency virus infection. Clin Infect Dis 2004;38:142-144.

70. Klencke BJ, Palefsky JM. Anal cancer: an HIV-associated cancer. Hematol Oncol Clin North Am 2003;17:859-872.

71. May T, Lewden C, Bonnet F, et al. Characteristics and causes of death among HIV-infected patients who had a good immunovirologic response under HAART: French National Survey "Mortalité 2000." 10th Conference on Retroviruses and Opportunistic Infections. February 10-14, 2003. Boston. Abstract 913.

72. Louie JK, Hsu LC, Osmond DH, et al. Trends in causes of death among persons with acquired immunodeficiency syndrome in the era of highly active antiretroviral therapy, San Francisco, 1994-1998. J Infect Dis 2002;186:1023-1027.

73. U.S. Preventive Services Task Force. Recommendations and rationale: screening for cervical cancer. www.ahrq.gov/clinic/3rduspstf/cervcan/cervcanrr.htm

74. de Sanjose S, Palefsky J. Cervical and anal HPV infections in HIV positive men and women. Virus Res 2002;89:201-211.

75. Chin-Hong PV, Palefsky JM. Natural history and clinical management of anal human papillomavirus disease in men and women infected with human immunodeficiency virus. Clin Infect Dis 2002;35:1127-1134.

76. Mary-Krause M, Cotte L, Simon A, et al. Increased risk of myocardial infarction with duration of protease inhibitor therapy in HIV-infected men. AIDS 2003;17:2479-2486.

77. Friis-M¿ller N, Sabin CA, Weber R, et al. Combination antiretroviral therapy and the risk of myocardial infarction. N Engl J Med 2003;349:1993-2003.

78. Fichtenbaum C, Lee D, Broderick W, et al. Selected hospital admission rates for patients who received HAART. 9th European AIDS Conference. October 25-29, 2003. Warsaw. Abstract 9.5/3.

79. National Cholesterol Education Program. Third report of the expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA 2001;285:2486-2497. www.nhlbi.nih.gov/guidelines/cholesterol/index.htm

80. Dubé MP, Stein JH, Aberg JA, et al. Guidelines for the evaluation and management of dyslipidemia in human immunodeficiency virus (HIV)-infected adults receiving antiretroviral therapy: recommendations of the HIV Medicine Association of the Infectious Disease Society of America and the Adult AIDS Clinical Trials Group. Clin Infect Dis 2003;37:613-27.

81. Salama C, Policar M, Cervera C. Knowledge of genotypic resistance mutations among providers of care to patients with human immunodeficiency virus. Clin Infect Dis 2003;36:101-104.

82. Salama C. Clinician knowledge and attitudes towards genotypic testing. XII International HIV Drug Resistance Workshop. Los Cabos, Mexico. Abstract 113.

83. Lorenzen T, Meyer P, Graefe K, et al. Radata -- expert advice conducted switch to antiretroviral therapy via compiling Internet database. 9th European AIDS Conference. October 25-29, 2003. Warsaw. Abstract 7.4/18.

84. Plattenberg A, Meyer P, Graefe K, et al. Resistance analysis helps to spare antiretroviral substances in pretreated patients -- subanalysis of the Radata project. 9th European AIDS Conference. October 25-29, 2003. Warsaw. Abstract 3.4/10.

85. Miller V, Phillips AN, Clotet B, et al. Association of virus load, CD4 cell count, and treatment with clinical progression in human immunodeficiency virus-infected patients with very low CD4 cell counts. J Infect Dis 2002;186:189-197.

86. Miller V, Mocroft A, Clotet B, et al. Association of viral load, CD4 cell count, and treatment with clinical progression in HIV patients with very low CD4 cell counts: the EuroSIDA cohort. 7th Conference on Retroviruses and Opportunistic Infections, San Francisco. January 30-February 2, 2000. San Francisco. Abstract 454.

87. Ledergerber B, Lundgren JD, Fusco GP, et al. Factors affecting CD4+ cell count slope in patients with stable viral load following three class virologic failure: the PLATO collaboration. 10th Conference on Retroviruses and Opportunistic Infections. February 10-14, 2003, Boston. Abstract 146lb.

88. Lundgren JD, Ledergerber B, Fusco GP, et al. Risk of death following triple class virological failure: the PLATO collaboration. 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy. September 14-17, 2003. Chicago. Abstract 450.