Home Search Abstracts Browse Sessions Program Committee E-mail Abstract Author View Session


Session 120 Poster Abstracts
Morbidity and Mortality of HIV-1 Infection
Monday, 1:30 - 3:30 pm
Poster Hall


871    
The Proportion of Deaths Due to HIV Infection among Persons with AIDS in New York City and San Francisco during 1987 to 1999
R Selik*1, T Singh2, and L Hsu3
1CDC, Atlanta, GA, USA; 2New York City Dept. of Hlth. and Mental Hygiene, NY, USA; and 3San Francisco Dept. of Publ. Hlth., CA, USA

Background:  The proportion of deaths of HIV-infected persons not caused by HIV has likely increased because of HAART.  To measure this increase, recent trends in underlying causes of death among persons reported with AIDS were examined, including persons whose death certificates did not mention HIV infection.

Methods:  AIDS cases reported to the health departments of New York City and San Francisco were matched to death certificate records for 1987 to 1999 from the health department’s vital statistics office or the National Death Index.  We used ICD-10 rules for selecting the underlying cause, applying them retroactively to deaths previously classified by ICD-9 rules.  In addition, we imputed HIV infection as the underlying cause for death certificates that had no mention of HIV if the initially selected underlying cause was an AIDS-indicative disease or HIV would have been selected if it had been mentioned on the death certificate.   We also expanded the criteria for selecting HIV infection as the underlying cause by attributing deaths to HIV infection if the initially selected underlying cause was cervical cancer, wasting/cachexia, dementia/encephalopathy, or meningitis. 

Results:  A total of 79,979 death certificate records were found for persons with AIDS who resided in New York City and San Francisco at diagnosis and died during 1987--1999.  Of these, 67,990 (85.0%) had HIV infection mentioned and selected as the underlying cause by ICD-10 rules.  Another 8073 (10.1%) had no mention of HIV infection, but 2523 of these were attributed to HIV infection by our imputation criteria.  Another 114 deaths were attributed to HIV infection by our expanded selection criteria.  Consequently, we attributed 70,627 (88.3%) of the deaths to HIV infection and 9352 (11.7%) to other underlying causes.  The most common other underlying cause was heart disease (1.9%).  The annual proportion of deaths for which HIV infection was not the underlying cause decreased from 10.6% in 1987 to 8.4% in 1995, and then increased to 22.9% in 1999, while the denominator (deaths of persons with AIDS) increased from 3992 in 1987 to 9217 in 1995, and then decreased to 2901 in 1999. 

Conclusions:  Almost a quarter of deaths of persons with AIDS have lately not been caused by HIV infection.  To avoid exaggerating the impact of HIV infection on mortality in the HAART era, the causes of death should be taken into account, rather than simply the number of deaths of persons with AIDS.

Keywords: causes of death; death certificates; acquired immunodeficiency syndrome