What our work means
By Benigno Rodriguez M.D. M.Sc. and Michael M. Lederman M.D.
Case Western Reserve University
Center for AIDS Research
Working with a large group of collaborators from across the nation, we recently published the results of a multicenter study designed to explore the role of the plasma HIV RNA level (also known as viral load) as a predictor of the rate of CD4 cell loss in individual HIV-infected persons. This publication, entitled "Predictive value of plasma HIV RNA level on rate of CD4 T cell decline in untreated HIV infection" (Rodriguez et al, JAMA 296:1498-1506, 2006), has generated extensive and sometimes impassionate discussion about the significance of HIV viremia and its role in HIV infection and disease progression, both among scientists and the general public. Positive as we believe cross-examination of scientific findings to be, we have learned with growing concern about interpretations of the work that are not only inaccurate, but misleading and potentially dangerous to HIV-infected persons everywhere. Thus, we are writing here to clarify the significance of this work, its implications for the role of HIV viral load measurement in clinical practice, and its meaning to persons living with HIV/AIDS.
Briefly, we used complex statistical modeling to calculate the estimated speed at which HIV-infected persons not receiving treatment for HIV would lose their CD4 cells over time, and then asked a simple question: based on a single measurement of those persons' viral loads, how well can one account for the variation in the rate of CD4 cell loss from one person to the next? To the surprise of many, the answer is "very poorly" - to the tune of only 4-6%. Most disturbing among all the interpretations of this finding, this has been taken by some to mean that our data raise doubts about HIV being the cause of AIDS; some have gone as far as to affirm that our results prove that it is not. As this is the most damaging of all the interpretations of our work, we will address it first.
There is absolutely no doubt that HIV is the cause of AIDS; far from challenging the veracity of this statement, our work further confirms it. This is easily appreciated from our initial analysis of the data, which shows that on average, individuals with higher viral loads tend to lose CD4 cells more rapidly that those with lower viral loads. There is no contradiction between this finding and our main message, because the overall trend among a group of subjects cannot be directly translated into a prediction of what will happen to a single individual within that group. Importantly, this finding replicates, rather than disputes, the substance of the seminal paper by Mellors et al (Ann Internal Med 1997; 126: 946-954), which demonstrated this almost 10 years ago. Thus, using our work to claim that those previous conclusions are invalid reveals either a combination of sloppy thinking, sloppy reading or malicious intent. We choose to believe that it is the first two.
What, then, are the implications of our work for understanding the biology of HIV infection? Twenty-five years after AIDS was first recognized, we have made enormous scientific and medical progress in learning about this disease and how to treat it, but great uncertainty still remains about how HIV infection causes the progressive immune deficiency that results in AIDS. For the past few years, a group of scientists (the Cleveland Immunopathogenesis Consortium), has been meeting regularly to join efforts to identify the mechanisms of HIV-induced immune deficiency; the work discussed herein is a product of this collaboration. Our results imply that although HIV infection drives the progressive immune deficiency of AIDS (as evidenced by the response to successful treatment with anti-HIV medicines, which decrease the viral load, increase CD4 cell numbers, and prevent or help resolve opportunistic infections, reflecting an improvement in immune function), there must be other intervening elements that cause progressive CD4 cell losses in HIV infection.
An oft-cited analogy posits that the clinical course of HIV infection can be thought of as a train approaching a broken bridge: the CD4 cell count is the distance that separates the train from certain doom, whereas the viral load is the speed at which the train is traveling towards that point. Expanding on this image, we propose that the train's fuel, rather than a single material, can be thought of as a mixture of combustibles, of which the number of viral particles in the blood (i.e., the viral load) is but one of the components. As the relative contribution of each component to the mixture changes, so does the efficiency of combustion and hence the power of the engine and the speed of the train. From this follows that were the train to run out of fuel, it would cease to move. This sine qua non in the equation is the presence of HIV in the system: no HIV, no AIDS. Thus, in two persons with the same amount of HIV in the blood, the efficiency of combustion and hence the speed of the train (rate of CD4 decline) may vary; that is precisely what our work shows. For the HIV-infected patient, this means that it is very difficult to predict what the pace of his or her CD4 cell decline will be just based upon measurement of the amount of HIV in blood. For this reason, more recent treatment guidelines have placed less emphasis on using HIV levels in blood to determine when to start treatment. Once antiviral treatment is started, however, it is critical to monitor the HIV levels in blood, because these levels remain the best indicator of the success of the treatment and the likelihood that its benefits will be sustained over time.
The next obvious question is, "What are the other elements in the fuel mixture that are driving the engine?" In the mid 1990's the late Janice Giorgi and others proposed that immune activation was a key element in driving the pace at which HIV disease progressed. Since then, many groups, including our own consortium, have been thinking along those very same lines as evidence that immune activation is an important element in driving CD4 T cell losses in chronic HIV infection accumulates. In chronic HIV infection, a variety of immune cells display telltale molecules that indicate that they have been activated, and as Dr. Giorgi pointed out years ago, the level of expression of these markers predicts the rate of HIV disease progression better than do blood levels of HIV. Equally important, a number of studies have shown the turnover of immune cells (that is, the process of division and death of these cells) is greatly accelerated in HIV infection, including cells that are not a primary target for HIV infection. The reason for this accelerated turnover is the subject of heated debate (The Push:Pull debate). Proponents of the pull theory believe that as HIV destroys immune cells, more cells are produced in order to make up for the losses, i.e., new cells are being pulled into division and expansion. Those who subscribe to the push model think that disorderly immune activation resulting from HIV infection pushes immune cells inappropriately into incomplete cell division and death. If you haven't guessed yet, we sit more with the pushers than with the pullers. So to our view, the key important question to ask now is how does HIV infection result in the systemic immune activation that drives CD4 T cell losses? We think that this will soon be illuminated. Stay tuned.
In summary then:
- HIV is the cause of AIDS
- In large groups of HIV infected persons who are not receiving antiretroviral therapy, those with higher levels of HIV in blood tend on average to lose CD4 cells from circulation faster than do those with lower levels of HIV in blood. Butů
- Levels of HIV in blood explain only a small portion of the variability in the rate at which CD4 T cells are lost. Therefore:
- For any one HIV infected person not receiving antiretroviral therapies it is difficult to predict the rate at which CD4 T cells will be lost.
- Expanded efforts to identify the other elements that drive CD4 cell losses in chronic HIV infection are needed.