Life expectancy in HIV

BMJ 2011; 343 doi: http://dx.doi.org/10.1136/bmj.d6015 (Published 11 October 2011) Cite this as: BMJ 2011;343:d6015
  1. Elena Losina, senior scientist1,
  2. Kenneth A Freedberg, director, program in HIV epidemiology and outcomes research2
  1. 1Brigham and Women’s Hospital, Boston, MA 02115, USA
  2. 2Massachusetts General Hospital, Boston, MA 02114, USA
  1. elosina@partners.org

Better, but not good enough

More than 33 million people are infected with HIV worldwide.1 Over the past 30 years, mortality from HIV and the life expectancy of people who are infected have improved dramatically. With major advances in biomedical research, increased awareness, and dedicated funding, HIV has been transformed from an untreatable and almost always fatal disease to a chronic one. For patients diagnosed promptly and treated with combination antiretroviral therapy (ART), life expectancy is now several decades.2 In the linked cohort study (doi:10.1136/bmj.d6016), May and colleagues estimate specific life expectancy for people in the United Kingdom with HIV undergoing treatment compared with life expectancy in the general population.3

Gains in life expectancy have increased steadily over time, with the availability of more effective and better tolerated regimens. But these gains have not been seen in everyone with HIV. Factors associated with worse outcomes include late presentation to healthcare services, suboptimal adherence to drugs, premature discontinuation of treatment, mental illness, and behavioural risk factors such as use of injected drugs and alcohol dependence.4

Data from the Joint United Nations Programme on HIV/AIDS (UNAIDS) suggest that more than 80 000 people are currently living with HIV in the UK, and about 25% of them are unaware of their infection.5 These people, their healthcare providers, and policy makers confront several key questions. How much life expectancy is lost as a result of HIV? How does the timing of the start of treatment affect life expectancy? Do losses in life expectancy as a result of HIV differ between men and women?

May and colleagues report estimates of life expectancy derived from a large cohort study of patients who started HIV treatment between 1996 and 2008 at some of the largest clinical centres in the UK. The authors suggest that between the periods 1996-9 and 2006-8, the life expectancy of an average 20 year old person infected with HIV increased from 30 to 46 years.

The authors also found that decreases in life expectancy as a result of HIV are greater in men than in women. They estimate that, for an average 20 year old man, HIV decreases life expectancy by 18.1 years; in contrast, a woman loses only 11.4 years. Why is the difference so large? Data from other countries show that women are likely to start treatment for HIV earlier than men, perhaps partly because women are often tested for HIV during pregnancy.6 7 Because earlier care is associated with better survival,8 this may explain the differences between men and women.

May and colleagues found greater reductions in life expectancy (more than 15 years lost) in those who start ART late (CD4 counts <100×106/L) rather than early (CD4 counts 200-350×106/L), providing more evidence in favour of earlier treatment. The presentation of information in terms of gains in life expectancy makes this important message easily understood by patients. For health related messages to be effective, people must perceive a problem as relevant and serious, and they should recognise that change provides clear gains. This study provides clinicians with the language to make these gains real.

May and colleagues’ study is an excellent example of a comprehensive analysis conducted on a well defined longitudinal cohort. However, the estimates should be interpreted within the boundaries of the data from which they are derived. The UK Collaborative HIV Cohort comprises data from referral centres. Although it is easier to conduct studies in high volume regional centres, not all HIV infected patients receive care in such settings. Patients of higher socioeconomic status are often over-represented in high volume clinics because those on low incomes and those in racial and ethnic minorities often receive care in lower volume centres.9 High volume referral centres are associated with better outcomes.10 11

The right censored nature of cohort data should also be taken into account. Participants are more likely to contribute early years on treatment, when mortality is lower, and to be censored later (because the follow-up period ends), when mortality is likely to rise. Because of the artificial right censoring that occurs when data are closed for analysis, people who started ART in 2006-8 had less follow-up time to contribute, which would also result in overestimation of recent survival and life expectancy.

Comparing life expectancy in people with HIV with that of the general population may misattribute losses to HIV that really come from other behavioural factors, such as smoking, substance misuse, and mental illness.6 12 Comparing life expectancy in those with and without HIV, but with similar risk factors, could shed light on this.

May and colleagues’ study serves as an urgent call to increase awareness of the effectiveness of current HIV treatments in patients and providers. In turn this should increase rates of routine HIV screening, with timely linkage to care and uninterrupted treatment. As these factors improve, the full benefits of treatment for all HIV infected people can be realised.


Cite this as: BMJ 2011;343:d6015


  • Research, doi:10.1136/bmj.d6016
  • Competing interests: All authors have completed the ICMJE uniform disclosure form at www.icmje.org/coi_disclosure.pdf (available on request from the corresponding author) and declare: no support from any organisation for the submitted work; no financial relationships with any organisations that might have an interest in the submitted work in the previous three years; no other relationships or activities that could appear to have influenced the submitted work.

  • Provenance and peer review: Commissioned; not externally peer reviewed.