Skip to content

The CD4 cell count at which combination antiretroviral therapy should be started is a central, unresolved issue in the care of HIV-1-infected patients.

— When to Start Consortium, “Timing of Initiation of Antiretroviral Therapy in AIDS-free HIV-1 infected Patients: A Collaborative Analysis of 18 HIV Cohort Studies.” Lancet 373;9672:1352-63, April 18, 2009.

Ever since the first anti-HIV drug was approved for prescription in 1987, there has been debate and controversy regarding when an HIV-positive person should start antiretroviral therapy (ART). Up until the mid-1990s, the question revolved around maximizing the limited benefits of treatment with one or two nucleoside analogue reverse transcriptase inhibitors (NRTIs) — such as AZT, ddI, ddC, d4T, 3TC, and the like — in temporarily staving off progression to AIDS. The advent of combination ART capable of prolonged — and potentially lifelong — suppression of HIV replication altered the landscape drastically, and now the focus is on the risks and benefits of earlier versus later treatment over decades. But for hundreds of thousands of HIV-positive people diagnosed at higher CD4 counts, this life-altering treatment decision has been fraught with uncertainty due to lack of the most reliable, rigorous evidence — that derived from well-designed, controlled, randomized clinical trials.

In the United States, the task of synthesizing the available evidence and making recommendations on how to use ART falls to a panel of experts who issue regularly updated guidelines under the aegis of the U.S. Department of Health and Human Services (DHHS). In December 2009 the DHHS guidelines panel revised its recommendation on when to start ART, raising the threshold from less than 350 CD4 cells to less than 500. The panel was split over the strength of this specific recommendation, with 55% endorsing it strongly and 45% moderately. Half of the panel also went even further, offering a moderate recommendation to start ART at any CD4 level, with the other half considering this approach optional. (The revised guidelines are available online.)

This change to the DHHS guidelines has generated controversy for a number of reasons. Most prominently, there is ongoing debate over the quality of the scientific evidence available to address the effect of starting ART at different CD4 cell count thresholds. The gold standard for evidence is the randomized controlled trial (RCT), in which individuals with very similar characteristics are randomly assigned to different interventions or strategies. In the absence of results from RCTs, data obtained from large cohorts of people with HIV in medical care is the most common secondary source of evidence. However, cohort studies are notoriously subject to a problem called confounding bias; a confounder is a factor associated with the outcome of interest that the study fails to capture. As an example, it could be that a member of a cohort who starts ART early has other factors associated with better health — such as access to private insurance and lower copays, or greater health-seeking behavior — and these factors could be more important to the study outcome than ART. Conversely, a cohort member who starts ART late may have risk factors for illness that the study fails to capture.

Without information on all potential confounders, as well as information on any negative aspects of HAART [highly active antiretroviral therapy] when initiated at higher CD4 cell counts, the benefit-to-risk ratio for early use of antiretroviral therapy remains unknown and we must await data from large RCTs [randomized clinical trials] before final conclusions can be reached.

— Caroline Sabin, “Cohort studies: to what extent can they inform treatment guidelines.” Curr Opin Infect Dis. 23;1:15-20, December 4, 2009.

There is virtually no evidence from RCTs of when to start ART among ART-naive individuals starting treatment with CD4 counts over 350/mm3. Most of the evidence cited in support of the new DHHS recommendations is derived not from RCTs but from two published meta-analyses of data from multiple large cohort studies, with additional support from smaller studies of HIV pathogenesis, particularly those describing harmful long-term effects of HIV-induced inflammation. A community sign-on letter addressed to the DHHS guidelines panel has expressed concern that the change could inadvertently make it more difficult to complete a critically important RCT that is being conducted by the INSIGHT network expressly to address the when-to-start question; this trial (known as START) is in a pilot phase and aims to compare initiation of ART at a CD4 cell count over 500 to deferral to a count of less than 350 cells (

Concerns about prevention have brought an additional wrinkle: for the first time, the DHHS guidelines note that suppression of viral load by ART is associated with a greatly reduced risk of transmitting HIV. This has caused some people to fear that the push to recommend earlier treatment is being made to prevent new infections. Despite the lack of conclusive evidence for clinical benefit to the individuals who will be initiating treatment for their own HIV infection.

Sorting though the tangle of issues now caught up in the when-to-start question is not easy. As a base line, there is widespread consensus that ART should be initiated when CD4 counts drop below 350, and this recommendation is supported by data from cohort studies and results from an RCT that were presented in 2009. The study, called CIPRA HT 001, enrolled 816 individuals in Haiti and compared starting ART with a CD4 count between 200 and 350 to deferring until the CD4 count fell below 200. An interim analysis revealed that deferral was associated with a significantly increased risk of illness and death and the differences were so stark (23 deaths in the deferred group versus 6 in the immediate group) that the trial was stopped by the Data Safety Monitoring Board.

At CD4 levels above 350, uncertainty intrudes. The only randomized data available are derived from a subset of participants in the Strategies for the Management of Antiretroviral therapy (SMART) trial, which was an evaluation of intermittent versus continuous ART. Two hundred and forty-nine people (out of a total of 5,472 participants) entered the study with >350 CD4 cells, having never taken ART. Of these, 131 were randomized to start ART immediately while 118 deferred until CD4 counts were lower than 250. Over an average follow-up of 18 months, there were seven cases of serious illness or death among people who deferred ART compared to two among people who started immediately. There was only one death, which occurred in the deferral group and was caused by cardiovascular disease. Because there were too few people in this subset for the results to reach statistical significance, the researchers conducted analyses that included individuals who had taken ART in the past but had been off for more than six months when they enrolled in SMART. With this group included, the difference in illnesses and deaths between the deferred and immediate groups became statistically significant, increasing the totals to 21 versus 6 events. The result was also significant if only those participants who had been off ART for at least a year were considered. However, there were some modest differences between participants who had never taken ART and the added group of those who had interrupted treatment. Members of the latter group were three years older on average and were more likely to have certain additional risk factors for illness (such as smoking) despite comparable CD4 cell counts (both current and the lowest levels ever reached). These factors may have exacerbated the risks associated with deferring ART. The authors of the paper describing these subset analyses from SMART (which was published in the Journal of Infectious Diseases in 2008) take pains to stress that the results are exploratory and need to be confirmed by an RCT.

The main evidence cited in support of the new DHHS recommendation to start at >350 CD4 cells comes from a large cohort study called NA-ACCORD (a “cohort of cohorts” that collates data from a number of smaller cohorts). In an influential paper published in the New England Journal of Medicine in 2009, researchers reported that cohort members who started ART at either >350 CD4 cells or >500 cells faced a significantly lower risk of death compared to those who deferred. It’s important to stress that, as with the SMART analysis described previously, the risk of death during follow-up was very low even among those who deferred. What the researchers emphasize is that the relative risk was significantly different, and this difference is expressed in terms that can easily mislead; for example, waiting until a CD4 count was below 500 was said to increase risk 94% compared to starting at above 500, and some people mistakenly interpreted this as suggesting that delaying ART meant death was a near certainty. However, in this context a 94% increase means roughly a doubling in risk, which for those with a CD4 count above 500 was relatively low (1.3 deaths per 100 person years, or approximately 13 deaths out of every 1, 000 people followed for a year).

The statistical approach used by the researchers did not allow an absolute death rate to be calculated for either of the deferral groups in the study. This approach is called inverse probability weighting and it is designed to address confounding factors that can affect comparisons of different treatment regimes using cohort data. Under this method, individuals are censored from the analysis if they deviate from the treatment regime being studied, and statistical modeling replaces the censored data with a “pseudoperson” for whom the outcome is calculated based on probabilities derived — by means that are not clear — from the same cohort. The confusing and opaque nature of this analytical method is one of the main concerns about the NA-ACCORD paper. The use of the inverse probability weighting technique by NA-ACCORD researchers has been questioned by Miguel Hernán and James Robins, the statisticians who originally developed it. Hernán has submitted a reanalysis of the same data for publication, but the article has yet to appear, and the extent to which it differs from the original results is not yet known.

The other major cohort study cited in the DHHS guidelines is the Antiretroviral Cohort Collaboration (ART-CC). The ART-CC’s analysis of 18 different cohorts supports the conclusions of the NA-ACCORD regarding starting at 350 CD4 cells but is equivocal at higher thresholds, prompting some debate between the two sets of researchers. Specifically, the ART-CC reported a reduced risk of reaching a composite endpoint of AIDS or death associated with starting between 350 and 450 CD4 cells compared to deferring to 250-350 CD4 cells, but when mortality was evaluated alone there was not a statistically significant difference. Furthermore, there were no significant differences associated with starting at CD4 thresholds over 450.

In addition to the cohort data, the guidelines articulate an important concern driving the shift to earlier ART: “later therapy may not repair damage associated with viral replication during early stages of infection. ” The SMART trial (which evaluated whether ART could be used intermittently to keep CD4 T-cell counts out of the danger zone for opportunistic infections, rather than continuously) brought risks associated with viral replication and the attendant inflammatory response to the fore in 2006. The results were unequivocal, showing that intermittent ART was associated with a doubling of the risk of serious illness and death when compared to continuous ART. Although absolute risk of illness and death was relatively low in both arms (around 3% and 1.5%, respectively), the difference was highly statistically significant. Differences in morbidity and mortality between the arms were also seen at all CD4 strata and not just the lowest. Most of the illnesses that occurred were not AIDS-defining events based on Centers for Disease Control and Prevention criteria but cardiovascular, kidney, and liver disease and cancers. The researchers who conducted SMART also collaborated with specialists from outside of the HIV research field to look for factors linked to illness and mortality in the trial. One such expert, Lewis Kuller, led a study that showed that biomarkers of inflammation (specifically the cytokine IL-6 and a marker linked to blood coagulation, D-dimer) were strongly linked to mortality in SMART, with statistical associations that dwarfed those Kuller had previously documented in non-HIV-infected elderly people. Other common risk factors were also predictive of mortality in SMART, including age, smoking, prior cardiovascular disease, co-infection with hepatitis B or C and baseline CD4 cell count, but the elevated inflammation associated with treatment interruption increased risk independently of any of these baseline variables. Important for the guidelines panel is that studies have found that the level of inflammation that persists after starting ART is associated with the CD4 cell count at starting; the lower the count, the higher the level of persistent inflammation.

Another related issue is the ongoing accumulation of worn-out or senescent T cells that accompanies uncontrolled HIV replication. Senescence is seen in both CD4 and CD8 T cells but is more prominent for CD8s, and research outside of the HIV field has shown that higher numbers of these cells are associated with illness, frailty, and earlier mortality among the very elderly. Senescent T cells are characterized by the loss of an important cell surface molecule called CD28, an inability to proliferate (copy themselves) in response to stimulation, production of high levels of proinflammatory cytokines, and a stubborn resistance to cell death (apoptosis). Recent studies in HIV have found that elevated numbers of senescent CD8 T cells are associated with more rapid disease progression and with inflammatory damage to blood vessels that presages cardiovascular disease. Relevant to the issue of when to start ART, it is as yet unclear whether there is much decline in numbers of senescent T cells after viral load is suppressed, suggesting that it may be better to intervene earlier in order to prevent their accumulation. The literature from research into the role of senescent T cells in aging indicates that they persist once present, raising the worry that if they are allowed to accumulate in a person with HIV they could be a harbinger of more rapid aging.

On the other side of the equation from the potential risks of deferring ART are the risks associated with the drugs themselves. Although the DHHS guidelines cite evidence that more recently marketed antiretrovirals have improved safety profiles when compared to those of earlier generations, they also acknowledge that long-term data are lacking. Somewhat reassuringly, neither the NA-ACCORD or ART-CC studies described earlier found evidence of harm associated with starting at high CD4 cell thresholds. However, as Caroline Sabin wrote in 2009, the only way to definitively characterize the risks and benefits of earlier treatment is via a randomized clinical trial.

Since the December 2009 guidelines update, conversations about when to start ART are likely happening in doctors’ offices around the country. Given the complexity of the available data and the uncertainty over the risk versus the benefit of starting early, making this treatment decision has not become any easier. As START trial sites open up around the country, enrolling in the study might be an option to consider. Besides contributing to the body of knowledge on HIV, being randomized to either the treatment or deferred arm of the study may be a way to turn a perplexing question into answers that will benefit countless people with HIV in the United States and around the globe.

For more information about the START trial, go to: To find a trial site near you, visit the website at:

This is the second in TAGLine’s 2010 coverage of ongoing changes in the standard of care for HIV treatment, following Mark Harrington’s “World Health Organization HIV Treatment Guidelines Evolve,” a review of the new World Health Organization adult and adolescent antiretroviral treatment guidelines.



Severe P, Pape J, Fitzgerald DW. A randomized clinical trial of early versus standard antiretroviral therapy for HIV-infected patients with a CD4 T cell count of 200-350 cells/ml (CIPRAHT001). Paper presented at: 49th Interscience Conference on Antimicrobial Agents and Chemotherapy;  Sep 12-15, 2009; San Francisco, CA. Abstract H1230c.

Strategies for Management of Antiretroviral Therapy (SMART) Study Group, Emery S, Neuhaus JA, Phillips AN, Babiker A, Cohen CJ, Gatell JM, Girard PM, Grund B, Law M, Losso MH, Palfreeman A, Wood R. Major clinical outcomes in antiretroviral therapy (ART)-naive participants and in those not receiving ART at baseline in the SMART study. J Infect Dis. 2008 Apr 15;197(8):1133-44.

Kitahata MM, Gange SJ, Abraham AG, et al. Effect of early versus deferred antiretroviral therapy for HIV on Survival. N Engl J Med. 2009;360(18):1815-1826.

Hernán MA, Robins JM. Early versus deferred antiretroviral therapy for HIV. N Engl J Med. 2009 Aug 20;361(8):822-3; author reply 823-4.

Sterne JA, May M, Costagliola D, et al. Timing of initiation of antiretroviral therapy in AIDS-free HIV-1-infected patients: a collaborative analysis of 18 HIV cohort studies. Lancet. 2009;373(9672):1352-1363.

Strategies for Management of Antiretroviral Therapy (SMART) Study Group, El-Sadr WM, Lundgren JD, Neaton JD, et al. CD4+ count-guided interruption of antiretroviral treatment. N Engl J Med. 2006 Nov 30;355(22):2283-96.

Kuller LH, Tracy R, Belloso W, et al. Inflammatory and coagulation biomarkers and mortality in patients with HIV infection. PLoS Med. 2008 Oct 21;5(10):e203.

Grund B, Neuhaus J, Phillips A; INSIGHT SMART Study Group. Relative risk of death in the SMART study. Lancet Infect Dis. 2009 Dec;9(12):724-5.

Larbi A, Franceschi C, Mazzatti D, Solana R, Wikby A, Pawelec G. Aging of the immune system as a prognostic factor for human longevity. Physiology (Bethesda). 2008;23:64-74.

Robbins GK, Spritzler JG, Chan ES, et al. Incomplete reconstitution of T cell subsets on combination antiretroviral therapy in the AIDS Clinical Trials Group protocol 384. Clin Infect Dis. 2009 Feb 1;48(3):350-61.

Cao W, Jamieson BD, Hultin LE, Hultin PM, Effros RB, Detels R. Premature aging of T cells is associated with faster HIV-1 disease progression. J Acquir Immune Defic Syndr. 2009 Feb 1;50(2):137-47.

Kaplan R, Sinclair E, Landay A, et al. T Cell Senescence and T Cell Activation Predict Carotid Atherosclerosis in HIV-infected Women. Paper #709, Seventeenth Conference on Retroviruses & Opportunistic Infections, San Francisco CA 2010.

Hunt P, Sinclair E, Epling L, et al. T Cell Senescence and Proliferation Defects Persist in Treated HIV-infected Individuals Maintaining Viral Suppression and Are Associated with Poor CD4+ T Cell Recovery. Paper #316, Seventeenth Conference on Retroviruses & Opportunistic Infections, San Francisco CA 2010.

Back To Top