March 28th, 2013

Poll: How Often Do You Measure CD4 Cell Counts?

Over in Clinical Infectious Diseases, a recent study pretty much nails the fact that routine measurement of CD4 cell counts in clinically stable patients is an all but useless exercise.  As summarized by Abbie Zuger in Journal Watch, here’s the key finding:

When patients with an unrelated cause for an alteration in CD4-cell count such as severe infection, chemotherapy, or interferon treatment were excluded from the analysis, not a single patient in any group had a dip in CD4 count below 200 cells/mm3 after 2 years of continuous virologic control.

So I’ve been singing this tune for a while (large nose-enhancing video here), and as a result have been trying for some time to get my stable patients to reduce the frequency of CD4 monitoring — or even, as I note in this editorial, give it up entirely!

And has this been a successful effort?  For some patients, yes, but for others it’s hopeless — they simply can’t understand that this test, which was the cornerstone of HIV monitoring for decades, now provides us with information that has no role in determining what we do therapeutically (provided the viral load remains suppressed).

In short, we order the test for emotional and sentimental reasons only — it’s reassuring to patients to hear that their CD4 cell count is stable, and we’ve been doing it for so long, why stop now. But are these good enough reasons? Before you answer in the affirmative, remember that unexplained drops in CD4 (which are not uncommon) have the opposite effect, and require frequent education about why the result won’t change our patients’ treatments.

So I ask you, providers of HIV care, the following burning question:

How often do you measure CD4 cell counts in your stable patients with long-term virologic suppression?

View Results

Loading ... Loading …

6 Responses to “Poll: How Often Do You Measure CD4 Cell Counts?”

  1. Dan Clinton says:

    Dr. Sax,
    I commend you for speaking up on the lack of value of CD4 counts in stable patients, but doesn’t such a significant shift call into question many of the fundamental tenants of the HIV-AIDS hypothesis. Why are CD4 counts fluctuating widely if viral load is stable and HIV causes the decline in CD4 counts seen in AIDS? How does HIV kill billions of T-cells it doesn’t infect since the original theory that HIV operates through a directly cytocidal mechanism has been accepted to be false since 1994? How is HIV more pathogenic after being neutralized by antibodies, in other words, how does it do more damage when founder in lesser concentration than it did upon initial infection? Why are there significant noncorrelations between viral load and CD4s? Why were so many important decisions based on data relating to CD4 counts? Doesn’t this call into question all the studies and treatment guidelines that justified treatment based on improving CD4 counts, and shouldn’t this make us much more skeptical of surrogate markers? Does it make any sense to classify patients with a CD4 count less than 200 as having AIDS although asymptomatic? Isn’t that a totally arbitrary line in the sand?

  2. Gary F. Spinner says:

    I started doing CD4 counts every six months about a year ago, with some consternation by my grants manager that our funders used CD4 count totals during a year as a measure of quality- expecting a minimum of 2 tests per year. Since sometimes every 6 months led to only one test a year for some patients who might have gotten two tests in 13 months, the grant CD4 “quality” measure showed a dip from the prior year. Since it is no skin off of my teeth to order the test with a viral load as I did in years past, I have agreed to return to the more frequent testing only to keep our grant quality measures high, much like measuring how many patients with CD4 counts below 200 are on Bactrim, another questionable measure of quality.

  3. Paul Sax says:

    An excellent example of how regulatory groups need to update their policies! In MA, our Drug Assistance Program requires CD4 as well as VL, for reasons that no longer make sense.

  4. Murugan Sankaranantham says:

    I will ask for CD4 counts with my patients once in a year only when they are doing clinically and physically well. When I am suspecting a clinical or immunological failure and also with in-patients suffering from OIs, I order for CD4 count as and when indicated without any time specification, sometimes even within a fortnight or a month.

  5. Maurizio Setti says:

    Dear Dr. Sax,
    CD4 count tells something about patient’s conditions even if it does not tell everything and it is not the only immune parameter to tell something.

    I agree that very seldom, if ever, we rely on CD4 count to take decisions about antiretroviral treatment; however we may still need to know such a value, and the other ones it brings along (CD4 percentage, CD8 number), to take decisions about several other steps of the HIV patient management (prophylactic drugs, coinfections and comorbidities management, vaccination strategies, etc.).

    I also agree that stable patients who do not suffer from intercurrent conditions such as severe infections, chemotherapy and interferon treatment – to which I would at least add severe liver disease or any condition of severe spleen enlargement – do not usually have significant
    CD4 cell count fluctuations, but such conditions may occur anyway. When they do, I want to know how far a CD4 cell count drop has to do with the new or the old illness.

    So, it may be fine to stop CD4 cell count in a stable patient with more than 700-800 cells over two or more years, and with a low number of
    CD8 cells, and with no comorbidity. It may not be fine to do so in a patient with 300-400 CD4 cells, a CD4/CD8 rate <1 and comorbidities, regardless if stable.

    To also respond to Dan Clinton comment, if CD4 cell count does not tell us everything, it does not mean it tells us nothing, it means we must find something else to tell us the rest of it.

  6. Dan Clinton says:

    I think what’s concerning is there remains no consensus about what a CD4 count actually means, just as there remains no vaccine predicted by 1987, no mechanism for how one virus could cause 26 previously-existing diseases all of which existed before HIV’s discovery and exist independent of HIV (AIDS), no internationally standardized definition of AIDS which itself is a unit of 27 and I have no idea why anyone would ever specialize in a unit of 27 or think that such a unit is particularly real, the predicted heterosexual pandemic never occurred, the “latency period” of the virus has been increased by more than a power of 10 meaning HIV was causing AIDS after 10 months in 1983 and after 10 years in 1993, there remains no internationally standardized diagnostic criteria for HIV (none of which actually detect a virion of HIV), to say HIV causes AIDS is to think in a circle giving AIDS is defined as any of 26 diseases + the designation HIV positive, creating a transient rise in CD4s was deemed an appropriate surrogate marker to prescribed 1500 MGs of AZT which helped to facilitate the death of 96% of patients when mortality decreased to 90% at 600 MG as shown by Fischl et all and then the Concorde showed the early AZT group had a 25% increased relative risk of death, and ARVs cause AIDS-defining disease but rather than recognize this a new term, Immune Reconstitution Syndrome, was created that hypothesizes the drugs improved the immune system which paradoxically gave the patient AIDS.
    Have you considered that you’re operating under a flawed hypothesis and there is a better explanation out there than one virus causing 26 previously-existing diseases?

Leave a Reply

Note: This is a moderated forum. By clicking on the "Submit Comment" button below, you agree to abide by the NEJM Journal Watch Terms of Use.

HIV Information: Author Paul Sax, M.D.

Paul E. Sax, MD


NEJM Journal Watch HIV/Aids Clinical Care

Biography | Disclosures | Summaries

Learn more about HIV and ID Observations.