Skip Navigation

Changing Spectrum of Mortality Due to Human Immunodeficiency Virus: Analysis of 260 Deaths during 1995–1999

  1. Hernan Valdez,
  2. Tanvir K. Chowdhry,
  3. Robert Asaad,
  4. Ian J. Woolley,
  5. Tracy Davis,
  6. Robin Davidson,
  7. Nele Beinker,
  8. Barbara M. Gripshover,
  9. Robert A. Salata,
  10. Grace McComsey,
  11. Sharon B. Weissman, and
  12. Michael M. Lederman
  1. Department of Medicine, Division of Infectious Diseases, and the Center for AIDS Research, Case Western Reserve University and University Hospitals of Cleveland, Cleveland
  1. Reprints or correspondence: Dr. Michael M. Lederman, 2061 Cornell Rm. 301 C, Cleveland, OH 44106 (lederman.michael{at}clevelandactu.org).

  1. Presented in part: 7th Conference on Retroviruses and Opportunistic Infections, San Francisco, February 2000.

 
Next Section

Abstract

We analyzed the deaths in an outpatient human immunodeficiency virus (HIV) care clinic at University Hospitals in Cleveland from January 1995 through December 1999. The number of annual deaths decreased progressively, from 112 in 1995 to 32 in 1999. The median final CD4+ cell count before death increased progressively from 10 cells/µL in 1995 to 90 cells/µL in 1999 (P<.01); 20%–25% of patients who died from 1997 through 1999 had plasma HIV RNA levels below detection limits. From 1995 through 1998, deaths due to infection, to end-stage acquired immune deficiency syndrome, and to malignancies decreased, whereas the proportion of deaths due to end-organ failures and of uncertain relationship to HIV infection increased. The spectrum of mortality in HIV disease has changed recently; although opportunistic infections cause death less frequently, deaths are occurring in people who have control of HIV replication and with some preservation of immune function. These observations underscore the need to monitor the etiologies of HIV-associated mortality and to better our understanding of the relationships among immune defenses, treatment-related toxicities, and end-organ failure in patients with HIV disease.

Improved prophylaxis for opportunistic infections and better treatments for HIV infection have contributed to a decrease in the rates morbidity and mortality due to HIV [1, 2]. Preliminary death statistics from the United States National Center for Health Statistics for a 12 month period that ended in June 1998 showed a 37% decrease in age-adjusted death due to HIV infection [3]. Palella et al. [2] described decreases in mortality from 29.4 deaths per 100 person-years in 1995 to 8.8 deaths per 100 person-years during the second quarter of 1997. Most of the benefit that was seen in that study occurred in patients with CD4+ T cell counts of <100 cells/µL. Chiasson et al. [4] found a 63% decrease in deaths due to HIV/AIDS in New York City from 1995 through 1997.

The purpose of this study was to analyze the numbers and causes of deaths in the John T. Carey Special Immunology Unit, University Hospitals of Cleveland, from 1 January 1995 through 31 December 1999, before and after the introduction of highly active antiretroviral therapy (HAART) into clinical practice.

Previous SectionNext Section

Methods

All recognized deaths and their dates were recorded. Approximately 6% of patients each year are lost to follow-up. Periodic queries made to the National Death Registry provided additional notice of deaths. Data regarding the patient's final CD4+ cell count, the patient's final plasma HIV RNA level, and the medications the patient was receiving at the time of death were obtained by review of the clinic's database, chart review, and hospital record review. For the purposes of this study, HAART was defined as concurrent treatment with ⩾3 antiretroviral medications.

The cause of death was ascertained by means of the following 4 methods: review of autopsy reports, review of the patients' hospital chart, interview with the patients' primary physicians, and review of death certificates. Discrepancies were resolved by use of a hierarchical method in which autopsy diagnoses were preferred over physician interviews, physician interviews were preferred over chart reviews, and chart reviews were preferred over diagnoses obtained from a review of death certificates. In 1995, 84% of the causes of death were ascertained by either review of the death certificate (for 56 [50%] of 122 patients who died) or review of medical records (for 38 [31%] of 122); in 1999, 25 (78%) of the causes of 32 deaths were ascertained by consultation with the patients' primary physicians (P<.001; χ2 test). Overall, only 11% of patients (28 of 255) underwent autopsy (table 1).

Table 1
View larger version:
Table 1

Method of ascertainment of cause of death for patients at the John T. Carey Special Immunology Unit, University Hospitals of Cleveland, 1995–1999.

Causes of death were classified by means of 2 schemes. In the first scheme, we sought to ascertain whether the deaths were directly related to HIV infection. Causes of death were classified as death due to an AIDS-defining illness [5], a recognized complication of HIV that was not an AIDS-defining illness, a recognized complication of HIV therapy (e.g., pancreatitis and abacavir hypersensitivity reaction), or a cause not directly related to HIV infection.

In the second scheme, we sought to describe the causes of death. Therefore, deaths were divided into the broad categories of “end-stage AIDS,” “infection,” “neoplasm,” “organ failure,” “trauma/accidental,” “unknown,” and >2 causes of death.” End-stage AIDS deaths were those deaths directly attributable to AIDS without an infection or neoplasm identified as the immediate cause of death; AIDS wasting syndrome was included in this category. End-organ failure was subclassified into the following causes: renal failure, liver failure, pancreatitis, cardiovascular failure, pulmonary failure, neurologic failure (including stroke and HIV encephalopathy), gastrointestinal bleeding not directly attributable to liver failure, and failure of >1 organ system. A hierarchy was used so that the underlying process was listed as the primary cause of death. For example, for a patient who died of aspiration pneumonia that was thought to be related to underlying CNS lymphoma, the cause of death was listed as “death due to CNS lymphoma.” If bacterial sepsis caused end-organ failure, sepsis was considered the cause of death.

Statistical analysis was performed by use of SPSS, version 9.0 for Windows (SPSS). Proportions were compared by use of the χ2 test and continuous variables were calculated by use of the Kruskal-Wallis test. All comparisons are 2-tailed; P<.05 was considered statistically significant.

Previous SectionNext Section

Results

A total of 260 patients died from January 1995 through December 1999. We were able to identify the cause of death in 255 (98%) of the 260 patients. The number of deaths that occurred each year and the demographic characteristics of the patients who died and of all patients followed in the clinic are shown in table 2. The number of deaths decreased from 112 in 1995 to 61 in 1996, 30 in 1997, and 20 in 1998. Although the number of deaths increased to 32 in 1999, the number of patients who were followed in the clinic also increased, from 548 in 1998 to 630 in 1999. Therefore, the frequency of death increased from 3.5 deaths per 100 patients followed to 4.8 deaths per 100 patients followed (P>.05; χ2 test). We have seen a nonsignificant increase in the proportion of women and ethnic minorities who have died during recent years. Although 55 (49%) of the 112 patients who died in 1995 were African-American, 22 (69%) of the 32 patients who died in 1999 were African-American. On a similar note, women accounted for 11 (10%) of 112 deaths in 1995, compared with 7 (22%) of 32 deaths in 1999. The percentage of patients who died without a prior opportunistic infection decreased significantly, from 73 (65%) of 112 patients in 1995 to 3 (15%) of 20 patients in 1998 (P<.001; χ2 test); however, it increased to 19 (59%) of 32 patients in 1999.

Table 2
View larger version:
Table 2

The demographics of the clinic and the characteristics of patients who died at the John T. Carey Special Immunology Unit, University Hospitals of Cleveland, 1995–1999.

The percentage of patients who were receiving HAART at the time of death increased significantly, from 7% (4 of 61 patients) in 1996 to 72% (23 of 32 patients) in 1999 (P<.001; χ2 test). The median final CD4+ T cell count recorded before death also increased significantly from 10 cells/µL in 1995 to 90 cells/µL in 1999 (P < .001 for comparison among groups by analysis of variance). The median duration from the time that the final CD4+ cell count was recorded to the time of death decreased significantly, from 128 days in 1995 to 56 days in 1999 (P<.001, by analysis of variance). In 1996, all of the 8 patients who had plasma HIV RNA levels analyzed before death had detectable levels. Thereafter, 21%–28% of all patients who died had final plasma HIV RNA levels that were below the limits of detection. Overall, 18 (22%) of the 83 patients who had a plasma HIV RNA level analyzed before death had <500 copies/mL of HIV RNA. The duration from the time that the final plasma HIV RNA level was recorded to the time of death ranged from 70 to 87 days.

Deaths were analyzed according to their relationship to HIV type 1 (HIV-1) infection. Deaths were defined as due to an AIDS-defining illness [5]; due to HIV-related immunosuppression or other HIV-related complications not included in the AIDS definition (i.e., cardiomyopathy or nephropathy); a recognized complication of antiretroviral therapy; not related to HIV infection; or of uncertain relationship to HIV infection. As is shown in table 3, the number of deaths due to AIDS-defining illnesses or related to HIV infection decreased dramatically, from 67 and 33, respectively, in 1995 to 10 and 7, respectively, in 1999. Likewise, the percentage of deaths due to an AIDS-defining illness decreased from 60% in 1995 to 31% of deaths in 1999 (P<.001; χ2 test; table 3). The percentage of deaths related to HIV infection but not due to an AIDS-defining condition remained relatively constant and accounted for 20%–30% of the deaths that occurred during the period of observation. In contrast, the 4 deaths attributed to treatment of HIV infection were identified after 1997. These deaths were attributed to pancreatitis (in 2 patients), liver failure (in 1), and abacavir hypersensitivity reaction (in 1). Finally, the number of deaths in this clinic that were deemed unrelated to HIV infection or of uncertain relationship to HIV infection remained relatively stable during the period of observation, whereas the proportion of deaths in these categories tended to increase while the proportion of deaths related to HIV infection decreased.

Table 3
View larger version:
Table 3

Relationship of cause of death to HIV infection for patients at the John T. Carey Special Immunology Unit, University Hospitals of Cleveland, 1995–1999.

We then categorized the causes of death as death due to end-stage AIDS (i.e., death attributable to AIDS or wasting syndrome, but with no clear opportunistic infection or neoplasm identified as immediate cause of death), infection, neoplasm, organ failure, trauma or accident, uncertain cause, and multiple causes (table 4). The number of deaths due to end-stage AIDS decreased from 25 in 1995 to 2 in 1999. Likewise, the number of deaths due to infection decreased from 48 in 1995 to 11 in 1999, and the number of deaths related to neoplasms decreased from 10 in 1995 to 1 in 1999. In this regard, there has been a decrease in the number of deaths due to CNS lymphomas; 4 and 3 deaths were due to this tumor in 1995 and 1996, respectively, but there was only 1 death due to this tumor in 1998 and 1 in 1999.

Table 4
View larger version:
Table 4

Causes of mortality for patients at the John T. Carey Special Immunology Unit, University Hospitals of Cleveland, 1995–1999.

The number and proportion of patients who died of end-stage AIDS (including AIDS wasting syndrome) decreased significantly (P<.001; χ2 test), from 25 (22%) of 112 patients in 1995 to 2 (6%) of 32 patients in 1999.

Although the numbers in each category are low, the number of deaths due to end-organ failure in the pre-HAART era (1995) and in the post-HAART era (1996–1999) did not change substantially, although there were 6 deaths due to liver disease in 1999 and only 2 in 1995. Nonetheless, the percentage of deaths related to end-organ failure increased significantly (P<.001; χ2 test), from 17% (19 of 112) in 1995 to 44% (14 of 32) in 1999.

Previous SectionNext Section

Discussion

As has been observed elsewhere, the advent of combination antiretroviral therapies has been associated with a dramatic decrease in the rate of mortality among persons with HIV infection [1, 2]. In our clinic, the rate of mortality decreased by ∼80% from 1995 through 1998. The rate of decrease stabilized in 1998 and 1999; although the number of deaths increased, the rates of death in 1998 and 1999 were not significantly different (P>.05; χ2 test). At present, we do not know whether this represents a fluctuation in death rates or the beginning of a trend of increasing mortality. The rising prevalence of treatment failure [6, 7], the apparent increase in resistance to antiretroviral drugs, and the additional complications of HIV infection or its treatment may contribute to an increasing risk of AIDS-associated mortality. Although an increasing proportion of the recent deaths in our clinic occurred among African-American patients, this increase was reflective of an increasing representation of African-American patients in our clinic, from 45% of patients in 1995 to 60% of patients in 1999.

There has been a shift in the proximate causes of death among patients in our clinic. In 1995, 48 infections accounted for 43% of deaths, whereas, by 1999, there were only 11 deaths that were directly attributable to infection (a decrease of ∼80%), and these deaths comprised 34% of the deaths that year. Likewise, the proportion of deaths due to an AIDS-defining illness decreased from 60% of the deaths in 1995 to fewer than one-third of the deaths in 1999. A comparison with the causes of death prevalent in 1995 shows that comparable numbers (but an increasing proportion) of deaths in 1999 were either due to a non-HIV-related illness or were of uncertain relationship to HIV infection. It is conceivable that some of these deaths were related to HIV infection or its treatment; therefore, monitoring the spectrum of these mortalities is particularly important, because they comprised an increasing proportion of deaths in this cohort.

The number of deaths attributed to end-organ failure remained fairly constant from 1995 through 1999; however, the proportion of deaths related to end-organ failures increased. By 1999, 14 (44%) of 32 deaths were due to end-organ failure. Among the mortalities attributed to end-organ failure in 1999, 4 of 14 were cardiovascular in origin and 6 of 14 were attributed to hepatic failure. Earlier studies have described a cardiomyopathy related to HIV infection [8–10]. The association of antiretroviral therapies with hyperlipidemias has also raised concern regarding accelerated atherosclerosis in patients with HIV disease; it is too soon and the number of events is too small to determine whether accelerated atherosclerosis played a role in the events that we report.

Recent reports have emphasized the occurrence of liver failure as a complication of HIV disease [11]. In some instances, this may be related to accelerated progression of hepatitis C virus infection [12–14]. Anecdotal reports and retrospective analyses have also implicated treatment with protease inhibitors as potential contributors to hepatic failure in patients with HIV disease [15–18].

In our clinic, the number of deaths related to malignancy decreased progressively, from 10 deaths in 1995 to 1 death in 1999. The incidence of Kaposi's sarcoma among people with AIDS started to diminish before HAART became available [19], and only 1 of the deaths in this series (in 1995) was attributed to Kaposi's sarcoma. Although there were relatively few deaths in our unit, the most apparent decrease was a decrease in the occurrence of CNS lymphoma, which has been reported elsewhere [20, 21]. In larger cohorts, there has not yet been a clear decrease in the incidence of other non-Hodgkin's lymphoma in people with AIDS [20–26]. Whether our more recent data will be reflective of a later trend of a decrease in the occurrence of these tumors, or whether the trend of a decrease in number of deaths related to these tumors reflects a better response to cancer chemotherapy in the era of HAART remains to be determined.

From 1997 through 1999, 16%–28% of deaths occurred among patients with final plasma HIV RNA levels of <500 copies/mL. Therefore, despite good control of HIV replication, some people with HIV disease remain at risk of dying. Likewise, the median final CD4+ T cell count recorded before the patient died increased from 10 cells/µL in 1996 to 90 cells/µL in 1999, which indicates that at least half of the people with HIV infection who died had some evidence of maintained or restored immune capacity. A more detailed analysis of these causes of death is underway.

The limitations of this study include uncertainties or bias regarding the cause of death and uncertainties regarding the frequencies of recent deaths. Only 11% of all patients who died received postmortem examinations. During 1995 and 1996, approximately half of the causes of death were determined by a review of death certificates, whereas, during 1998 and 1999, the attending physician determined 85% and 78% of causes of deaths, respectively. Therefore, some degree of time-dependent bias may limit the accuracy of our data. However, we think this is not likely, because in virtually all instances, the physicians who cared for our patients had also been the physicians who completed the death certificates and the medical records from which causes of death were determined. Because data from the National Death Registry were included only for 1995 and 1996, mortalities for 1997–1999 also might be underestimated. This potential bias confounds most retrospective studies.

Previous SectionNext Section

Conclusion

This retrospective analysis of 260 deaths that occurred during a 5-year period demonstrates a 70%–80% decrease in the number of deaths that occurred between 1995 and 1998 and 1999. Much of the decrease in mortality is attributable to a decrease in deaths attributed to end-stage AIDS (including wasting syndrome) and HIV-related opportunistic infections. In recent years, approximately one-fifth to one-quarter of deaths occurred in patients with good control of HIV replication, and an increasing median CD4+ cell count in these patients suggests that a greater proportion of these deaths are now less frequently due to HIV-induced immune deficiency than were deaths in the past. Our results underscore the need both to monitor the etiologies of HIV-associated mortality and to improve our understanding of the relationship among immune defenses, treatment-related toxicities, and end-organ failure in patients with HIV disease.

Previous SectionNext Section

Acknowledgments

We thank the staff of the John T. Carey Special Immunology Unit for their assistance in organizing and retrieving data used for the generation of this report.

Previous SectionNext Section

Footnotes

  • Financial support: National Institutes of Health (grant Al 36219).

  • Received July 12, 2000.
  • Revision received September 20, 2000.
Previous Section
 

References

  1. ↵
    1. Hogg RS,
    2. O'Shaughnessy MV,
    3. Gataric N,
    4. et al
    . Decline in deaths from AIDS due to new antiretrovirals [letter]. Lancet 1997;349:1294.
    MedlineWeb of Science
  2. ↵
    1. Palella FJ Jr,
    2. Delaney KM,
    3. Moorman AC,
    4. et al
    . Declining morbidity and mortality among patients with advanced human immunodeficiency virus infection. HIV Outpatient Study Investigators. N Engl J Med 1998;338:853-60.
    CrossRefMedlineWeb of Science
  3. ↵
    1. Smith BL,
    2. Martin JA,
    3. Ventura SJ
    . Births and deaths: preliminary data for July 1997–June 1998. Natl Vital Stat Rep 1999;47:1-32.
    Medline
  4. ↵
    1. Chiasson MA,
    2. Berenson L,
    3. Li W,
    4. et al
    . Declining HIV/AIDS mortality in New York City. J Acquir Immune Defic Syndr 1999;21:59-64.
    MedlineWeb of Science
  5. ↵
    1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults. MMWR Morb Mortal Wkly Rep 1992;41(RR-17):1-19.
    Medline
  6. ↵
    1. Deeks SG,
    2. Hecht FM,
    3. Swanson M,
    4. et al
    . HIV RNA and CD4 cell count response to protease inhibitor therapy in an urban AIDS clinic: response to both initial and salvage therapy. AIDS 1999;13:F35-43.
    CrossRefMedlineWeb of Science
  7. ↵
    1. Valdez H,
    2. Lederman MM,
    3. Woolley I,
    4. et al
    . Human immunodeficiency virus 1 protease inhibitors in clinical practice: predictors of virological outcome. Arch Intern Med 1999;159:1771-6.
    Abstract/FREE Full Text
  8. ↵
    1. Cohen IS,
    2. Anderson DW,
    3. Virmani R,
    4. et al
    . Congestive cardiomyopathy in association with the acquired immunodeficiency syndrome. N Engl J Med 1986;315:628-30.
    MedlineWeb of Science
    1. Kaminski HJ,
    2. Katzman M,
    3. Wiest PM,
    4. et al
    . Cardiomyopathy associated with the acquired immune deficiency syndrome. J Acquir Immune Defic Syndr 1988;1:105-10.
    MedlineWeb of Science
  10. ↵
    1. Barbaro G,
    2. Di Lorenzo G,
    3. Grisorio B,
    4. Barbarini G
    . Incidence of dilated cardiomyopathy and detection of HIV in myocardial cells of HIV-positive patients. Gruppo Italiano per lo Studio Cardiologico dei Pazienti Affeti da AIDS. N Engl J Med 1998;339:1093-9.
    CrossRefMedlineWeb of Science
  11. ↵
    1. McGovern B,
    2. Bica I,
    3. Stone D,
    4. Snydman D
    . Increasing mortality from end-stage liver disease secondary to hepatitis C in patients with human immunodeficiency virus infection [abstract 205]. Clin Infect Dis 1999;29:998.
  12. ↵
    1. Darby SC,
    2. Ewart DW,
    3. Giangrande PL,
    4. et al
    . Mortality from liver cancer and liver disease in haemophilic men and boys in UK given blood products contaminated with hepatitis C. UK Haemophilia Centre Directors' Organisation. Lancet 1997;350:1425-31.
    CrossRefMedlineWeb of Science
    1. Benhamou Y,
    2. Bochet M,
    3. Di Martino V,
    4. et al
    . Liver fibrosis progression in human immunodeficiency virus and hepatitis C virus coinfected patients. The Multivirc Group. Hepatology 1999;30:1054-8.
    CrossRefMedlineWeb of Science
  14. ↵
    1. Dieterich DT
    . Hepatitis C virus and human immunodeficiency virus: clinical issues in coinfection. Am J Med 1999;107:79-84.
  15. ↵
    1. Vergis E,
    2. Paterson DL,
    3. Singh N
    . Indinavir-associated hepatitis in patients with advanced HIV infection. Int J STD AIDS 1998;9:53.
    FREE Full Text
    1. Picard O,
    2. Rosmorduc O,
    3. Cabane J
    . Hepatotoxicity associated with ritonavir [letter]. Ann Intern Med 1998;129:670-1.
    FREE Full Text
    1. Brau N,
    2. Leaf HL,
    3. Wieczorek RL,
    4. Margolis DM
    . Severe hepatitis in three AIDS patients treated with indinavir [letter; comment]. Lancet 1997;349:924-5.
    MedlineWeb of Science
  18. ↵
    1. Arribas JR,
    2. Ibanez C,
    3. Ruiz-Antoran B,
    4. et al
    . Acute hepatitis in HIV-infected patients during ritonavir treatment [letter]. AIDS 1998;12:1722-4.
    MedlineWeb of Science
  19. ↵
    1. Mocroft A,
    2. Sabin CA,
    3. Youle M,
    4. et al
    . Changes in AIDS-defining illnesses in a London Clinic, 1987–1998. J Acquir Immune Defic Syndr 1999;21:401-7.
    CrossRefMedlineWeb of Science
  20. ↵
    1. Jones JL,
    2. Hanson DL,
    3. Dworkin MS,
    4. Ward JW,
    5. Jaffe HW
    . Effect of antiretroviral therapy on recent trends in selected cancers among HIV-infected persons. Adult/Adolescent Spectrum of HIV Disease Project Group. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S1-7.
    Medline
  21. ↵
    1. Sparano JA,
    2. Anand K,
    3. Desai J,
    4. Mitnick RJ,
    5. Kalkut GE,
    6. Hanau LH
    . Effect of highly active antiretroviral therapy on the incidence of HIV-associated malignancies at an urban medical center. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S18-22.
    MedlineWeb of Science
    1. Jacobson LP,
    2. Yamashita TE,
    3. Detels R,
    4. et al
    . Impact of potent antiretroviral therapy on the incidence of Kaposi's sarcoma and non-Hodgkin's lymphomas among HIV-1-infected individuals. Multicenter AIDS Cohort Study. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S34-41.
    MedlineWeb of Science
    1. Buchbinder SP,
    2. Holmberg SD,
    3. Scheer S,
    4. Colfax G,
    5. O'Malley P,
    6. Vittinghoff E
    . Combination antiretroviral therapy and incidence of AIDS-related malignancies. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S23-6.
    MedlineWeb of Science
    1. Grulich AE
    . AIDS-associated non-Hodgkin's lymphoma in the era of highly active antiretroviral therapy. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S27-30.
    MedlineWeb of Science
    1. Mueller N
    . Overview of the epidemiology of malignancy in immune deficiency. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S5-10.
  26. ↵
    1. Rabkin CS,
    2. Testa MA,
    3. Huang J,
    4. Von Roenn JH
    . Kaposi's sarcoma and non-Hodgkin's lymphoma incidence trends in AIDS Clinical Trial Group study participants. J Acquir Immune Defic Syndr 1999;21(Suppl 1):S31-3.
    MedlineWeb of Science
| Table of Contents

This Article

  1. Clin Infect Dis. (2001) 32 (10): 1487-1493. doi: 10.1086/320164
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most