BHIVA: One-in-three UK HIV deaths 'not directly related to HIV'

Who reported data?

BHIVA's mortality audit consisted of responses from 133 clinical centres throughout the United Kingdom, 80% of which were outside of the London region. These centres were of various sizes, with 19% serving 1-50 HIV patients, 23% 51-100, 20% 101-200, 21% 201-500, and 17% more than 500 patients.

A total of 40 centres reported no deaths among their adult HIV patients during the audit period. Consequently, 89 centres submitted case note review data for 397 deaths among adults with HIV. Ten patients died outside of the audit period (October 2004-September 2005) and were excluded from analysis. Although the date of death was missing for a further eight patients, these were included in the analysis. Therefore, a total of 387 deaths were analysed.

One-in-three deaths that occurred in HIV-positive individuals in the United Kingdom between 2004 and 2005 were not directly related to HIV, according to the final results of the British HIV Association (BHIVA) mortality audit presented at BHIVA's Autumn conference, held last week in London. The audit also found that cancers related to HIV, as well as those traditionally not related to HIV, accounted for more deaths than any other cause. Other major non-"classical" AIDS-defining specific causes of death included liver disease due to hepatitis B/C co-infection and/or alcohol, and cardiovascular disease.

Comparing data and demographics

The Health Protection Agency's SOPHID (Survey of Prevalent HIV Infections Diagnosed) data reported 467 deaths in HIV-positive individuals in the 2004 calendar year. The number of annual deaths in HIV-positive individuals reported to the HPA has consistently been around this level since 1998, dropping substantially from around 1,800 in 1995. However, since the number of HIV-positive individuals is constantly increasing, the proportion of deaths is actually falling. Consequently, only around 1% of all diagnosed HIV-positive individuals actually died in 2004.

The BHIVA audit found that three-quarters of all deaths occurred in men, and that 57% of all deaths occurred in people of white ethnicity, 33% in people of black-African ethnicity and 2% in people of black-Caribbean ethnicity. In comparison, SOPHID's 2004 data (which include all age groups in England, Wales and Northern Ireland, whereas BHIVA's data include adults in the UK) report that two-thirds of people seeking HIV-related care were men, 52% were white, 38% black-African and 3% black-Caribbean.

Half had CD4 cell counts above 200 cells/mm³

CD4 and viral loads in the last six months of life were not uniformly poor. In fact, around half of all people who died had CD4 cell counts above 200 cells/mm³ and 30% had 'undetectable' viral loads. Consistent with these data is the finding that one-third of deaths were not considered to be directly linked to HIV infection.

These non-HIV -related deaths comprised:

• 30 due to non-HIV-related cancers (7.8% of all deaths)
• 22 due to end-stage liver disease (5.7%)
• 17 due to cardiovascular disease (4.4%)
• 7 due to suicide (1.8%)
• 7 due to bacterial sepsis (1.8%)
• 6 due to accident or injury, including one homicide (1.6%)
• 4 due to accidental injection drug overdose (1.0%)
• 1 due to end-stage kidney disease (0.3%)

Another 29 deaths (7.5%) were either due to other non-HIV-related causes or the cause was not stated.

Other causes and factors

Cardiovascular disease was the immediate cause of death for 25 (6.5%) patients, of which 17 were considered not to be directly related to HIV. The majority (18 or 4.6% of all deaths) of deaths due to cardiovascular disease were due to coronary artery disease, with two due to HIV-related pulmonary hypertension and three due to other heart muscle diseases.

Liver disease due to hepatitis B/C co-infection and/or alcohol accounted for a further 6% of all deaths.

The audit found that deaths due to adverse reactions to therapy, immune reconstitution inflammatory syndrome (IRIS) or multi-drug resistant HIV - each accounting for less than 3% of all deaths - were relatively rare. Of note, there were three cases of lactic acidosis, which is associated with mitochondrial toxicity due to NRTI antiretroviral therapy.

Patient factors leading to death included poor adherence (26, or 6.7% of all deaths); declining treatment (18, or 4.7% of all deaths); and, in two cases, poor HIV clinic attendance or lack of regular care.

Cancers accounted for one-in-six deaths

The single largest immediate cause of death was bacterial sepsis, accounting for one-in-eight of all reported deaths. However when HIV-related and non-HIV-related cancers were combined they accounted for one-in-six of all reported deaths.

These cancer-related deaths included:

• 29 due to lymphoma* (7.3% of all deaths)
• 6 due to liver carcinoma, of which two were reported as liver disease rather than cancer) (1.5%)
• 6 due lung cancer (1.5%)
• 3 due to anal cancer* (0.75%)
• 2 due to adenocarcinoma (glandular cancer) (0.75%)
• 2 due to kidney cancer (0.75%)
• 2 due to oesophageal cancer (0.75%)
• 2 due to penile cancer (0.75%)
• 2 due to prostate cancer (0.75%)
• 1 each due to bladder, bowel, breast, cervical*, Merkel cell, multiple myeloma, pancreatic cancers
• 5 were not known or not stated, but one was considered to be directly related to HIV

* Considered to be directly related to HIV

Some limitations

A limitation of the audit is that the data were collected via HIV clinicians, who may not have known of all deaths. In fact, the audit uncovered some issues regarding communication and record-keeping which may have lead to under-ascertainment of deaths, including undiagnosed HIV disease in community or general medical settings as well as deaths in the community due to non-HIV-related causes such as suicide or drug overdose.

Conclusion

Along with late diagnosis, causes not directly related to HIV account for the majority of deaths in adults with HIV. Deaths due to adverse reactions to HIV therapy and to running out of options with multi-drug resistant HIV were, according to Professor Sebastian Lucas of St. Thomas’ Hospital, London, who presented the data with BHIVA’s President, Professor Margaret Johnson, reassuringly rare.