SEARCH

SEARCH BY CITATION

Keywords:

  • causes of death;
  • HIV/AIDS;
  • mortality rates;
  • non-HIV related deaths

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Objectives

To examine changes over a 2-year period in both the mortality rate and the causes of death in a geographically defined HIV-infected population.

Methods

A database search of primary care information for the dates and causes of death for all patients documented with HIV infection and living in Southern Alberta between 1984 and 2003 was undertaken. Sociodemographic and clinical characteristics were obtained. Causes of death were then individually confirmed by reviewing the patients' hospital charts, autopsy reports, or death certificates and coded using the International Classification of Diseases, 9th Revisions. AIDS deaths were reconciled with Public Health Reports. The time span was divided into pre-highly active antiretroviral therapy (HAART) (1984–1996) and current HAART (1997–2003) periods.

Results

Between 1984 and 2003, there were 560 deaths in the 1987 individuals living with HIV infection in Southern Alberta. Of these, 436 deaths (78%) occurred pre-HAART and 124 (22%) in the current HAART period. The crude mortality rate declined from 117 deaths per 1000 patient-years pre-HAART to 24 in the current HAART period. In the pre-HAART era, 90% of all deaths were AIDS related whereas only 67% were AIDS related in the current HAART era. The leading causes of AIDS deaths were AIDS multiple causes (31%), Mycobacterium avium complex (18%), Pneumocystis pneumonia (10%) and non-Hodgkin's lymphoma (7%). The proportion of non-AIDS related deaths increased from 7% pre-HAART to 32% in the current HAART era. Accidental deaths, including drug overdose (29%), suicide (7%) and violence (3%), hepatic disease (19%), non-AIDS related malignancies (19%), and cardiovascular disease (16%) accounted for the majority of non-AIDS related deaths. No deaths directly caused by drug toxicity were found. Overall, 21% of patients who died were antiretroviral (ARV)-naïve. A total of 14% of patients dying from AIDS were ARV-naïve in contrast to 35% dying from non-HIV related conditions. Of all those dying from AIDS, 23% died<3 months after their initial diagnosis, reflecting late presentation. In the current HAART era, 87% of patients who died from AIDS were extensively treated, reflecting HAART treatment failures due mostly to multiclass drug resistance (42%), inexorable disease progression despite ARV (32%), lack of ability or interest to be maintained on a lifelong HAART programme (21%) and, rarely, drug intolerance (<1%).

Conclusions

Deaths from AIDS-related causes have decreased significantly, but deaths from non-AIDS related conditions have increased, both as an absolute number of deaths and as a proportion of all deaths in HIV-infected patients. The increasing age of the HIV population, and the increased mean CD4 count, increased proportion of intravenous drug users, increased hepatitis B virus and hepatitis C virus coinfection rate, and increased history of smoking seen in our population also influenced the mortality rate and causes of death. These factors must also be considered in projecting future trends in mortality of an HIV-infected population.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Over the past 8 years, the overall AIDS-related mortality rate within the HIV-infected population in developed countries has dropped significantly following the discovery and subsequent extensive use of highly active antiretroviral therapy (HAART) [1–9]. Increasing numbers of deaths of HIV-infected individuals, however, are being reported for reasons other than AIDS, such as hepatitis B and C virus (HBV and HCV) infection, drug overdoses, and malignancies [10–13]. In addition, AIDS-related deaths are still being reported for diverse reasons, including failure of or late access to care, antiretroviral toxicities and failure of HAART secondary to drug-resistant virus or advanced disease [14–17].

We wished to review the death rate for all patients with HIV infection living in our region between 1984 and 2003, and examine the causes of death linked to their ‘clinical histories’.

Patients and methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study population and site

In Southern Alberta, HIV care is centralized to the Southern Alberta Clinic (SAC) programme, which provides exclusive out-patient HIV care at one of four sites to all infected individuals living in the region. Patients are continuously followed from their initial HIV diagnosis or move into the region until they move out of the region or die. Since 1983, extensive clinical and sociodemographic data have been collected on each new clinic patient. Patients not accessing regular HIV care also came to our attention when they presented with terminal disease or when an autopsy report confirmed AIDS, which is a reportable condition. The data thus represent a complete regional population of HIV-infected individuals over the entire AIDS era (1984–2003) within a universal health care system.

Methodology

A search of the database for all dates and causes of death was linked to sociodemographic data (age, gender, ethnicity, HIV exposure category, and history of smoking) and clinical data (CD4 count, AIDS diagnosis, time since HIV diagnosis, antiretroviral use, and HBV or HCV coinfection). CD4 counts were used only if taken within 6 months of death. Time of follow-up was based on the number of months a patient was followed at SAC from their initial visit to either the time they died or the time they moved out of the region. Antiretroviral (ARV)-naïve patients were defined as those not having received any antiretroviral drugs at any time during the course of their illness. HAART was defined as treatment with three or more drugs concurrently, usually from two or more classes.

Cause of death was individually confirmed by reviewing the patient's discharge summary hospital chart, autopsy report, or death certificate and coded using the International Classification of Diseases, Ninth Revisions (CD-9 codes). When the cause of death was problematic, a secondary review was undertaken by an experienced clinician (MJG) to confirm the correct coding. AIDS deaths, which are reportable in Canada, were reconciled with Public Health Records.

Deaths were categorized as either AIDS-related or non-AIDS-related deaths. AIDS-related deaths were those where the primary cause of death was from one of the 21 AIDS-defining illnesses. When multiple concurrent AIDS diagnoses contributed to the death, the ‘AIDS multiple’ category was used. Non-AIDS-related deaths were placed into one of five categories: (i) cardiovascular (heart/stroke), (ii) hepatic, (iii) non-AIDS-defining malignancy, (iv) suicide/overdose/violence, and (v) other.

The location where the patient died, when available, was categorized as (i) hospital, (ii) home, (iii) hospice, (iv) long-term care facility, (v) out of province, or (vi) unknown.

We divided the time span into two periods, pre-HAART and HAART, using 31 December 1996 as the binary cut-off date.

Statistical comparisons were made using the χ2 likelihood ratio or Fisher's exact text, where appropriate, and Student's t-test for comparison of mean values. All analyses were two-tailed with P< 0.05 considered significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Mortality rate

Between January 1984 and December 2003, there were 560 deaths of HIV-infected individuals living in Southern Alberta. The three patients who died from AIDS before receiving any HIV care were included in the analysis. During this time, 1987 HIV-infected patients received care at some point at the SAC, with a median follow-up time of 42 months [interquartile range (IQR) 19–77]. A total of 436 deaths (78%) occurred in the pre-HAART period and 124 (22%) in the current HAART period. The crude mortality rate declined from 117 deaths/1000 patient-years pre-HAART to 24 deaths/1000 patient-years in the current HAART period (Fig. 1).

image

Figure 1.  Crude mortality rate (1984 to 2003)=66 per 1000 patient years. Pre-highly active antiretroviral therapy (HAART) (1984-1996)=117.0. Current HAART (1997-2003)=24.0.

Download figure to PowerPoint

Causes of death

The cause of death was well documented for 98% of patients. During the entire study period, 85% of all deaths were attributable to AIDS-related conditions (Fig. 2). However, whereas pre-HAART 90% of all deaths were AIDS-related, only 67% were directly attributable to AIDS in the current HAART period (Figs 2a and b).

image

Figure 2.  Proportion of all documented deaths (1985-2003) by cause of death (n=560).

Download figure to PowerPoint

AIDS-related deaths

The median CD4 count taken within 6 months of death for those dying from AIDS was 15 cells/μL (IQR 6–42). There was little difference in median CD4 count between the pre-HAART period (14 cells/μL, IQR 6–45) and the current HAART period (23 cells/μL, IQR 7–100).

The leading cause of AIDS-related death was ‘AIDS multiple’ (30%), followed by individual diseases such as Mycobacterium avium complex (MAC) infection (18%), Pneumocystiscarinii pneumonia (PCP) (10%), non-Hodgkin lymphoma (7%), Kaposi's sarcoma (7%), encephalopathy (6%), wasting syndrome (5%) and cytomegalovirus (CMV) disease (5%), as shown in Fig. 3. Most of the individual AIDS diseases decreased in proportion from the pre-HAART era to the current HAART era (e.g. MAC, 18 to 13%; PCP, 10 to 6%; Kaposi's sarcoma, 8 to to 2%); however, non-Hodgkin's lymphoma and encephalopathy increased from 6 to 11% and from 5 to 7%, respectively. ‘AIDS multiple’ increased from 29 to 42% of all AIDS-related deaths, reflecting increased complexity of care.

image

Figure 3.  Individual causes of AIDS deaths (n=482).

Download figure to PowerPoint

Non-AIDS related deaths

Non-AIDS related deaths increased both as an absolute number and as a proportion of all patient deaths (Fig. 1). In the pre-HAART, non-AIDS related deaths accounted for 7% of all deaths, whereas in the current HAART period they account for 32%. During the study period, accidental death including drug overdose (29%), suicide (7%) and violence (3%), non-HIV related malignancies (19%), hepatic disease (16%) and cardiovascular conditions (16%) were the leading causes of non-AIDS related deaths (Fig. 4). Although actual numbers remained small, hepatic disease (10 to 20%) and accidental death (35 to 43%) increased in proportion. Patients who died from non-AIDS related conditions had significantly higher CD4 counts than those who died from AIDS. Their median CD4 value was 264 cells/μL (IQR 78–434), in contrast to 15 cells/μL (IQR 6–42) for the AIDS patients. This reflects their better immunological health.

image

Figure 4.  Proportions of non-AIDS deaths (n=75).

Download figure to PowerPoint

Accidental deaths, primarily from drug overdose, increased from 2.2 to 17% of all deaths from the pre-HAART period to the current HAART period. Deaths from non-HIV malignancies increased from < 1 to 7%. Deaths from liver disease increased from < 1 to 8.4% of all deaths, reflecting an increasing prevalence of HBV and HCV coinfection. Cardiovascular deaths remained stable and low at 2.4%.

We did not find any deaths that appeared to be attributable, directly or indirectly, to the well-described antiretroviral drug toxicities such as lactic acidosis, pancreatitis, Stevens Johnson syndrome, hepatic necrosis, renal failure or abacavir hypersensitivity.

Antiretroviral (ARV) use

ARV-naïve patients

Overall, of patients who died, 21% were ARV-naïve. Of patients dying from AIDS, 14% were ARV-naïve, in contrast to 35% of those dying from non-HIV-related causes. Twenty-three per cent of ARV-naïve patients dying from AIDS died within 3 months of their HIV diagnosis, reflecting late presentation of their HIV illness. Of all patients dying of AIDS, only three patients had never received any HIV care prior to death. Fifteen per cent of patients living with HIV for longer than 3 months died from AIDS pre-HAART without ever having received ARV drugs, but only 9% had not received ARV drugs in the current HAART era.

ARV-experienced patients

Overall, 95% of HIV patients dying from AIDS had received ARV for 3 or more consecutive months at some time during their illness. From 1990 to 1995 (i.e. late pre-HAART), 86% of patients who died from AIDS had been on ARV drugs at some point, while in the current HAART period 87% of patients dying from AIDS were extensively HAART-treated. These latter deaths were HAART treatment failures and mostly were attributable to multiclass drug resistance (42%), progressive disease despite HAART (32%), or lack of ability or interest to be maintained on a lifelong HAART regimen (21%), with only a small number unable to tolerate HAART (see Table 1).

Table 1.   Most likely contributing factor for the cause of death for extensively HAART-treated HIV-infected patients in the current HAART period (1998–2003)
 N%
Multi-class drug resistance2342
Progressive disease despite HAART1732
Lack of ability or interest to be on a lifelong HAART regimen1120
Unable to tolerate HAART24
Unknown12
Total54100

Sociodemographic characteristics

Death rates in an HIV-infected population may be influenced not only by HIV disease progression but also by the characteristics of the active population [18–22]. We examined changes in five common factors, both HIV- and non-HIV related, that may be associated with increased mortality rates: age, CD4 count, intravenous drug use (IDU), HCV coinfection, and smoking (see Table 2).

Table 2.   Selected sociodemographic and clinical characteristics of the entire HIV-infected population (i.e. all HIV-infected patients followed) compared with those of HIV-infected patients who died during the pre-HAART (1984–1996) and current HAART (1997–2003) periods
 Pre-HAART periodTotal populationCurrent HAART periodP- value
Total populationPatients who diedPatients who died
  1. na, not available.

Age
% > 40 years294253420.05
HIV exposure category
% intravenous drug use161121320.01
CD4 count
% > 200 cells/μL581083180.001
Coinfection
% HIV/HCVnana20630.001
Cigarette smoking
% with history of smoking411757550.05

During the pre-HAART period, 28% of all HIV-infected patients were >40 years old; this proportion increased to 51% in the current HAART period. In contrast, 42% of HIV-infected patients dying during both the pre-HAART and current HAART periods were >40 years old (P<0.01).

From the preHAART to the HAART period, the proportion of all HIV-infected patients with a CD4 count of >200 cells/μL increased from 58 to 83%, indicating better overall health of the population. Not unexpectedly, only 10% of patients dying in the pre-HAART period and 18% of those dying in the current HAART period had CD4 counts >200 cells/μL.

In our population, IDU as the sole exposure category for HIV infection increased from 16 to 21%. For patients who died, this proportion increased from 11% pre-HAART to 32% in the current HAART era.

The prevalence of HCV in our population in the pre-HAART period is unknown, but in the current HAART period, when HCV testing became widely available, 63% of patients who died were HIV/HCV coinfected. This is significantly higher than the HCV prevalence rate of 20% in the general HIV-infected population (P< 0.001).

The percentage of individuals with a history of cigarette smoking has remained significantly higher in our HIV-infected population than in the general population of Alberta. During the pre-HAART period, 38% of all Albertans reported smoking regularly. In the HIV-infected population for this period the proportion was slightly higher at 41%. In the current HAART period, cigarette smoking for all Albertans has decreased to 23%, whereas for HIV-infected individuals the proportion has increased to 57%. Somewhat surprisingly, of those HIV-infected patients dying during the pre-HAART era, only 17% reported smoking use; in the current HAART era, 55% of HIV-infected patients who died smoked.

Site of death

The site of death was documented in 286 (54%) cases. Overall, 42% of deaths occurred in hospital, 41% at home, and 15% in a hospice or long-term care facility. In the pre-HAART period, 50% of deaths occurred in hospital, while this proportion has decreased to 25% in the HAART period. Deaths at home have increased from 38 to 49%. Deaths at a hospice or long-term facility have increased from 11% pre-HAART to 24% in the HAART period.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We have tracked the sociodemographic, clinical, and mortality characteristics for our regional HIV-infected population over most of the AIDS epidemic (1984–2003). Our population has become more diverse in terms of age (i.e. older), HIV exposure category [(i.e. more injecting drug users (IDUs)], and presence of concurrent diseases (such as HCV), but it has also become healthier as proxied by a higher mean and median CD4 count. HAART has had a dramatic effect in significantly reducing mortality from AIDS, as seen when comparing deaths in the pre-HAART and current HAART periods.

In our population, the mortality rate in the current HAART period is one-fifth the rate in the pre-HAART period. AIDS-related deaths continue to decrease during the current HAART period, whereas non-HIV-related deaths increase. In the current HAART period, non-AIDS-related deaths account for over 33% of all deaths in our HIV-infected population. Deaths from accidental causes, cardiovascular disease and liver disease, although low in absolute numbers, have all increased as a proportion of all deaths. The majority of accidental deaths are from drug overdoses, reflecting the increase in IDUs within the regional HIV-infected population.

The increase in cardiovascular disease and liver disease is more problematic. The high rate of smoking in our population may increase the risk for cardiovascular disease. The precise combinations of ARV drugs that may be contributing to cardiovascular death in HIV-infected patients remain difficult to assess, but they may also be a contributing factor. We determined by careful chart review that the small but significant increase in deaths from liver disease in our population was unlikely to be caused by antiretroviral drug toxicities, as has been described, but this still remains an important consideration.

Low CD4 count (i.e. <200 cells/μL) and experience of an AIDS-defining illness remain highly correlated with death, although in the current HAART period a significantly greater number of patients die with very high CD4 counts and no AIDS-defining illnesses than in the pre-HAART period. The majority of these individuals die from non-HIV related conditions. HCV infection contributes increasingly to the overall number of deaths in the current HAART period, where approximately 66% of all patients who died were coinfected. Approximately, the same proportions of patients in the pre-HAART and HAART periods were ARV-naïve prior to death. Although half of the patients who died between 1996 and 2003 had never received HAART, the majority of these individuals did not receive HAART because they died shortly after their HIV diagnosis or they were noncompliant with their care.

Whereas older, Caucasian males with men having sex with men (MSM) as their likely HIV exposure category were over-represented in deaths in the HIV-infected population over the entire AIDS era (P<0.05), the group most at risk of dying in the current HAART period are young male IDUs, reflecting the sociodemographic shift in the regional population.

These data support our observation that non-AIDS related deaths remain an important cause of mortality in an HIV-infected population, and that the underlying characteristics of the population may determine the mortality rate of that cohort. The lack of competing risks as a result of HIV-infected patients not dying from AIDS-related diseases such as MAC or PCP is also a contributing factor in increasing rates of deaths from non-AIDS-related conditions.

Conclusions

HAART has increased the overall health of the majority of HIV-infected patients in our population by increasing, stabilizing, or delaying the decline of CD4 counts. The overall mortality rate has decreased, with deaths from HIV-related causes declining dramatically, but deaths from non-HIV related causes have increased. This shift probably reflects the improvement in the immunological health of the clinic population from the pre-HAART to the current HAART period, as well as the changing demographics of the patients served. HAART has increased survival and therefore extended the mean age of our population. This increased longevity, however, has increased the patient's risk of dying from other conditions not directly related to HIV infection (e.g. cancer, suicide or hepatic disease) or side effects perhaps related to long-term use of ARV drugs. Non-compliance or non-adherence and late diagnoses of HIV disease are also ongoing contributing factors to death in the current HAART period, as are treatment failures with resistant viruses.

Changing demographics, including an increase in intravenous drug use, have also contributed to a significant increase in non-AIDS related deaths within the HIV-infected community. At this stage of the HIV epidemic it is difficult to know if the trends described here are a reflection of variegated effects of the HIV virus itself or simply a reflection of a severely HIV-impacted population becoming ‘normalized’ towards the mean mortality rates seen in the Canadian population. Continued long-term study of an entire large HIV population will allow this problem to be addressed. As HIV disease shifts from an acute to a chronic condition, however, we should expect to see these trends continue.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

We wish to acknowledge the contributions made by Jim Jewett in the collection of the data, and Mike Henry in assisting with the data analysis. The study was partially supported by a grant from the Canadian Institutes of Health Research.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  • 1
    Pallella FJ Jr, Delancy KM, Moorman AC et al. Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998; 338: 853860.
  • 2
    Jain MK, Skiest DJ, Cloud JW, Jain CL, Burn D, Berggren RE. Changes in mortality related to human immunodeficiency virus infection: comparative analysis of inpatient deaths in 1995 and in 1999–2000. Clin Inf Dis (CID) 2003; 36: 10301038.
  • 3
    Schanzer DL. HIV/AIDS mortality in trends in Canada, 1987–1998. Can J Public Health 2003; 94: 135139.
  • 4
    Mocroft A, Ledergerber B, Katlama C et al. Decline in the AIDS and death rates in the EuroSIDA study: an observational study. Lancet 2003; 362: 2229.
  • 5
    Van Sighem AI, Van De Wiel MA, Ghani AC et al. Mortality and progression to AIDS after starting highly active antiretroviral therapy. AIDS 2003; 17: 22272236.
  • 6
    Hogg RS, O'Shaughnessy MV, Gataric N et al. Decline in deaths from AIDS due to new antiretrovirals. Lancet 1997; 349: 1294.
  • 7
    Porter K, Babiker A, Bhaskaran K et al. Determinants of survival following HIV-1 seroconversion after the introduction of HAART. Lancet 2003; 362: 12671274.
  • 8
    Messeri P, Lee G, Abramson DM, Aidala A, Chiasson MA, Jessop DJ. Antiretroviral therapy and declining AIDS mortality in New York City. Med Care 2003; 41: 512521.
  • 9
    Schanzer DL. Trends in HIV/AIDS mortality in Canada, 1987–1998. Can J Public Health 2003, 94: 135139.
  • 10
    Rodriquez BV, Valdez H, Salata H et al. Changes in causes of death in a cohort of HIV-infected individuals: analysis of the last 6 years. Programs and Abstracts of the 9th Conference on Retroviruses and Opportunistic Infections. Seattle, WA, February 2002 [Abstract 755-W].
  • 11
    Palella FJ Jr, Baker R, Moorman AC, Chmiel J, Wood K. Holmberg SD and the HOPS Investigators. Mortality and morbidity in the HAART era: changing causes of death and disease in the HIV Outpatient Study. Programs and Abstracts of the 11th Conference on Retroviruses and Opportunistic Infections. San Francisco, CA, February 2004 [Abstract 872].
  • 12
    Rimland D, Guest J. Mortality in the HIV Atlanta VA Cohort Study: a 22-year analysis. Programs and Abstracts of the 11th Conference on Retroviruses and Opportunistic Infections. San Francisco, CA, February 2004 [Abstract 873].
  • 13
    Bonnet F, Lewden C, May D et al. Malignancies-related causes of death of HIV-infected patients in the era of highly active antiretroviral therapy. Programs and Abstracts of the 11th Conference on Retroviruses and Opportunistic Infections. San Francisco, CA, February 2004 [Abstract 875].
  • 14
    Wolfe MI, Hanson DL, Selik DL, Swerdlow R. Deaths from non-AIDS related diseases have increased as a proportion of deaths of HIV-infected persons since the advent of HAART. Programs and Abstracts of the 9th Conference on Retroviruses and Opportunistic Infections. Seattle, WA, February 2002 [Abstract 14].
  • 15
    Jambroes MD, Dieleman J, Van De Wiel J et al. Changing mortality in an HIV-infected population on HAART. Programs and Abstracts of the 9th Conference on Retroviruses and Opportunistic Infections. Seattle, WA, February 2002 [Abstract 754-W].
  • 16
    Selik R, Singh T, Hsu L. The proportion of deaths due to HIV infection among persons with AIDS in New York City and San Francisco during 1987 and 1999. Programs and Abstracts of the 11th Conference on Retroviruses and Opportunistic Infections. San Francisco, CA, February 2004 [Abstract 871].
  • 17
    Wood E, Montaner JSG, Tyndall MW, Schecheter MT, O'Shaughnessy MV, Hogg RS. Prevalence and correlates of untreated human immunodeficiency virus type 1 infection among persons who have died in the era of modern antiretroviral therapy. J Inf Dis (JID) 2003; 188: 11641170.
  • 18
    Sabin C. The changing clinical epidemiology of AIDS in the highly active antiretroviral therapy era. AIDS 2002; 16: S61S68.
  • 19
    Knodel J, Van Landingham M. The impact of the AIDS epidemic on older persons. AIDS 2002; 16: S77S83.
  • 20
    Haydon GH, Mutimer DJ. Hepatitis B and C virus infections in the immune compromised. Curr Opin Infect Dis 2003; 16: 473479.
  • 21
    Clarke CA, Glaser SL. Epidemiologic trends in HIV-associated lymphomas. Curr Opin Infect Dis 2001; 13: 354359.
  • 22
    Chiao EY, Krown SE. Update on non-acquired immunodefiency syndrome-defining malignancies. Curr Opin Infect Dis 2003; 15: 389397.