when to start meds in HIV patients

A recent large multinational  study (START trial: Strategic Timing of Antiretroviral Therapy) from 215 sites in 35 countries assessed the effect of randomizing patients with CD4 counts >500 to immediate antiretroviral therapy (ART) vs deferring therapy until the CD4 counts decreased to 350 or the development of AIDS or other condition requiring ART (eg pregnancy) --see hiv start therapy early NEJM 2015 in dropbox, or DOI: 10.1056/NEJMoa1506816. This study was funded by the National Institute of Allergy and Infectious Diseases and others. details:

--4685 patients (median age 36, 27% women) followed 3.0 years. The trial was stopped early because of clear benefit in the interim analysis, and all patients were offered ART. 32.8% of the patients were from Europe/Israel, 25.1% from South America/Mexico, 21.3%  from Africa, 10.8% from North America, 7.6% from Asia. 55% of those developing HIV were MSM, 38% heterosexual contact, 1.4% injection drug use, 5.2% from blood products

--median HIV viral load at the study onset was 12,759 copies/ml and CD4 was 651. 

--primary endpoint was composite of any serious AIDS-related event (though they excluded esophageal candidiasis, nonfatal HSV infection), serious non-AIDS-related event (eg CAD, ESRD, liver disease, non-AIDS defining cancer, non-AIDS related death) and death from any cause

--results:

    --for those in the deferred group, the median CD4 count at initiation of ART was 408.

    --the most common drugs used were tenofovir (89%), emtricitabine (89%) and efavirenz (though efavirenz was used in 73% in the immediate treatment group and 51% in the deferred. the other drugs were the same in both groups). viral suppression was achieved in 98% in both groups. drug initiation in the immediate ART group increased the CD4 dramatically to >900 by 24 months.

    --primary endpoint occurred in 42 patients (1.8%) in the immediate ART group and 96 patients (4.1%) in the deferred group, for HR 0.43 (0.30-0.62, p<0.001). 71% of the events in the deferred treatment group occurred prior to starting ART

            ​--serious AIDS-related events had HR 0.28 (0.15-0.50, p<0.001), largely by reduction in rates of TB, Kaposi's, malignant lymphomas

            --serious non-AIDS-related events had HR 0.61 (0.38-0.97, p=0.04), largely by reduction in non-AIDS-defining cancers.

            --death from any cause: no significant difference between the groups

            --the most common events in the immediate-initiation group and deferred group were cardiovascular disease (29% and 15%, respectively); non-AIDS-defining cancer (21% and 19%), and TB (14% and 20%). Most of the TB cases occurred in Africa and most of the cancer and cardiovascular disease in Europe/Israel, Australia, and the US

    --more than 2/3 of the primary endpoints occurred in patients with CD4>500!!!! (in fact for the immediate ART group 94% of the time the CD4 count was >500, and 72% of the time in the deferred group)

    --secondary endpoints: no difference in grade 4 events (potentially life-threatening symptomatic events not attributable to AIDS) or in unscheduled hospital admissions

    ​--subgroup analysis basically showed that all benefited from early ART (by age, sex, race, geographic region, baseline CD4 or viral load, smoking status, Framingham CAD risk score)

so, a few observations.

--this study adds solid evidence that it is beneficial to start ART​ early and should be offered to all infected patients. This has been the recommendation in the US (see http://gmodestmedblogs.blogspot.com/2015/04/updated-hiv-guidelines-2015.html  ), though a large part of the rationale was to prevent transmission of the virus (treatment as prevention). This study shows important efficacy for the people themselves who are infected, even those with "normal" CD4 counts (it is pretty striking that more than 2/3 of the clinical problems in the delayed start group occurred at CD4 counts >500)

--the international implications of the study are profound, both clinically and cost-wise.  I suspect that the World Health Organization will endorse the "treating all" approach, again raising their threshold to treat from the current CD4 counts of <500, leading to millions more individuals qualifying for therapy. The TEMPRANO study (see doi/full/10.1056/NEJMoa1507198​) simultaneously published in New Engl J Med also found that early initiation of ART and of anti-TB therapy in the Ivory Coast each led to 35-44% decrease in risk of death or severe HIV-related illness, even in those with baseline CD4 counts >500.

--but, in reviewing the results buried in the supplemental material, it was interesting to note the HIV-related events disproportionately found in the delayed treatment group: pulmonary TB in 16 vs 6, Kaposi's sarcoma in 11 vs 1, PCP (pneumocycstis jirovecii) in 5 vs 1, non-Hodgkins lymphoma (NHL) in 9 vs 2. To me, this is pretty striking. PCP, for example, much more commonly happens when the CD4 is below 200, NHL typically with long-standing infection and severe immunocompromise. These unexpected results raise the question (also raised in blog http://gmodestmedblogs.blogspot.com/2015/04/placebo-genetics-and-placebome.html  ) that it might have been useful to have matched non-HIV infected controls included. Are there other infections/exposures which could contribute to the clinical outcomes (both AIDS-related and otherwise) found here, perhaps related to the different regions of the world studied? Although the subgroup analysis did not find different overall results in the different regions, were there differences in the types of outcomes? (they did mention in broad strokes that there were more HIV-related deaths in the delayed ART group in Africa and more non-HIV related in Europe, but did not reveal the region-specific differences in detail). Having a non-HIV infected control group might help sort this out.