Starting antiretroviral therapy earlier, before the development of symptoms, is the most likely way to reduce the high death rates after treatment initiation seen in people with HIV in resource-limited settings, two large cohort analyses show. The studies also show that the major disadvantage of starting treatment late – an increased risk of death – may persist for some years, burdening already overstretched health systems with illness that could be avoided by earlier treatment.
The findings, presented last week at the Sixteenth Conference on Retroviruses and Opportunistic Infections in Montreal, are likely to strengthen the case for a stronger global recommendation that people with HIV should start treatment when the CD4 cell count falls below 350 cells/mm3, wherever resources permit.
Current World Health Organization guidelines endorse treatment for anyone with a CD4 count below 200, and advise clinicians to 'consider' treatment in anyone with a CD4 count between 200 and 350 cells/mm3, with the aim that treatment should start before the CD4 count falls below 200 cells/ mm3.
Current national guidelines in many countries in Asia and sub-Saharan Africa continue to emphasise treatment for those with CD4 counts below 200 cells/mm3, due to concerns about the cost of treatment and the capacity to deliver it to larger numbers of patients (see HIV & AIDS Treatment in Practice June 2008 for further discussion of these constraints and how they are being addressed at country level).
However there is accumulating evidence that, due to late diagnosis, many people are continuing to start HIV treatment very late – often at CD4 counts below 50 cells/mm3. These individuals are much more likely to die or develop serious illnesses after starting treatment and, as Dr Stephen Lawn of the University of Cape Town pointed out to the conference, “Our hospitals are chock-a-block and just can’t cope” as a consequence of the failure to identify patients with HIV earlier and to begin treatment earlier.
A systematic review of published studies by Stephen Lawn and colleagues published last year shows death rates of between 8% and 26% among cohorts starting treatment late.
The two studies presented at CROI 2009 provide compelling evidence of the costs of late treatment initiation, for people with HIV in terms of lives lost, and for health systems in terms of the burden of avoidable disease occurring as a result of late identification and treatment of people with HIV infection.
The consequences of late treatment initiation in Gugulethu, South Africa
Stephen Lawn presented results from patients receiving antiretroviral treatment through a community-based programme in Gugulethu, a township near Cape Town. The study looked at the risk of death not only on the basis of the CD4 count at the time treatment was started, but on the basis of the updated CD4 count at specific time-points during the four-year follow-up period.
Follow-up data were available for 2423 people who had initiated antiretroviral therapy, observed for 3155 person-years. The study population’s median baseline cell count was 105 cells/mm3; CD4 cell count levels were subsequently measured at four-month intervals.
The cumulative mortality after 48 months of follow-up was 13.2%, lower than many cohorts in Africa, but among individuals who started treatment with a CD4 count below 100 and an AIDS-related illness, the 48-month cumulative mortality was 24.8%.
There were 197 deaths, with the following associations between updated CD4 cell count levels and mortality rate ratios: 0 to 49 cells/mm3, 11.6; 50 to 99 cells/mm3, 4.9; 100 to 199 cells/mm3, 2.6; 200 to 299 cells/mm3, 1.7; 300 to 399 cells/mm3, 1.5; 400 to 499 cells/mm3, 1.4; and 500 cells/mm3 or more, 1.0.
When the researchers calculated person-time within updated CD4 cell count strata, they found high mortality during the first year of treatment to be related to the large proportion of person-time spent at less than 200 CD4 cells/mm3. People with less than 100 CD4 cells/mm3 had higher cumulative mortality estimates at one and four years than those whose baseline CD4 cell counts were higher (one year: 11.6% vs 5.2% mortality; four-year: 16.7% vs 9.5% mortality).
The researchers attribute this to person-time spent at low CD4 cell counts, and conclude, “National HIV programmes in resource-limited settings should be designed to minimise the time that patients spend with CD4 counts [below] 200 cells/mm3 both before and during [use of antiretroviral therapy].”
Dr Lawn said that health systems have two options: either continue in "firefighting" mode, treating AIDS-related illnesses as they arise in people who start treatment late, or treat people earlier and reduce the burden of mortality and morbidity caused by HIV. But, he warned, “firefighting is very difficult and time-consuming and takes up a huge amount of resources.”
Mortality in people with HIV compared to background population levels
The other study compared the mortality rates of antiretroviral-treated people in four sub-Saharan African countries to all-cause mortality in the corresponding populations. HIV-related mortality data were drawn from antiretroviral treatment programmes in Ivory Coast, Malawi, South Africa and Zimbabwe. The researchers used estimates from the World Health Organization (WHO) Global Burden of Disease project to calculate the expected numbers of non-HIV-related deaths in the relevant populations.
Data on clinical stage of HIV disease were available for 13,249 HIV-positive people with 14,695 person-years of follow-up; 85% had advanced disease when they began taking antiretroviral therapy. One thousand, one hundred and seventy-seven deaths occurred during 14,695 person-years of follow-up, a cumulative mortality of 11.7%.
They found that people with advanced HIV disease (3) had a mortality rate 500 times higher than the background rate in the local population, and among those who started treatment with a CD4 count below 25 cells/mm3 the mortality rate remained 47 times higher than the background rate in the local population after two years of follow-up. Among those who started treatment without symptoms and a CD4 count above 200 the mortality rate remained slightly elevated compared to the general population, but was substantially lower than that seen in those who started treatment much later (1.24 to 3.4-fold higher than the general population).
Presenting the results, Martin Brinkhof of the University of Berne, concluded that while there is greater mortality among HIV-positive people taking antiretroviral therapy than in the general population, “for some patients the excess is moderate and mortality reaches that of the general population in the second year of [antiretroviral therapy].” They add that more timely initiation of antiretroviral therapy might prevent much of the excess mortality.
Brinkhof M et al. Mortality of HIV-infected patients starting ART: comparisons with the general population in Southern Africa. Sixteenth Conference on Retroviruses and Opportunistic Infections, Montreal, abstract 141, 2009.
Lawn S et al. Changing mortality risk associated with CD4 cell response to long-term ART: Sub-Saharan Africa. Sixteenth Conference on Retroviruses and Opportunistic Infections, Montreal, abstract 140, 2009.