Print this page    •   Back to Web version of article

When to Start Antiretroviral Therapy in HIV-Infected Patients
An Interview With Joel Gallant, M.D., M.P.H., and Paul Sax, M.D.

By Myles Helfand

July 18, 2012

HIV therapy is a rapidly changing field. One might hope that more than 30 years of intimate experience with a virus might leave us with a firm sense of precisely how to treat it. But HIV has largely eluded easy answers, and antiretroviral therapy continues to evolve dramatically almost from year to year.

As a result, seemingly simple questions -- such as "When do I start my patient on antiretroviral therapy?" -- don't have simple answers. In this interview, I was honored to discuss this most fundamental of HIV treatment questions, along with other related issues, with two of the world's leading HIV specialists: Joel Gallant, M.D., M.P.H., a professor of medicine and epidemiology at the Johns Hopkins University School of Medicine and the associate director of the Johns Hopkins AIDS Service; and Paul Sax, M.D., the clinical director of the HIV Program and the Division of Infectious Diseases at Brigham and Women's Hospital, as well as the editor-in-chief of Journal Watch HIV/AIDS Clinical Care.

Both men are currently members of the panel of experts that periodically revises the bible of HIV treatment in the U.S.: the Department of Health and Human Services (HHS) antiretroviral therapy guidelines for HIV-infected adults and adolescents.

Table of Contents

Why We've Become So Aggressive About HIV Treatment

Myles Helfand: It has not been a ridiculous number of years since U.S. HHS HIV treatment guidelines set 200 CD4+ cells as the threshold for starting treatment. And now, here we are: As of March of 2012, the guidelines have been updated to include a recommendation that people start treatment when their CD4+ cell count is above 500. What has led us to where we are now, in terms of the aggressiveness of when to start antiretroviral therapy?

Joel Gallant: Part of it is that the regimens and medications we have are so much better than the ones we had when the cutoff was 200. We used to expect inevitable toxicity with our antiretrovirals. That's no longer the case. We can usually treat with a single, daily dose of medications that are very well tolerated and have pretty minimal long-term toxicity.


In addition, I think we've come a long way from the days when we thought that the only bad thing that happened to you with HIV was that you had a low CD4 count, and that you got opportunistic infections and malignancies. We now know that HIV is a disease of inflammation; of immune activation that, over the long term, can cause medical consequences beyond those of the traditional list of opportunistic infections. These things are more independent of CD4 counts -- and more, perhaps, driven by viral replication. So there's a benefit to suppressing viral replication, regardless of what your CD4 count is.

And then, finally -- and maybe what tipped the scales over to this approach -- has been the results of HPTN 052, showing that antiretroviral therapy for positive patients is the most effective form of prevention that we have. And so when you combine the benefits to the individual who is infected, and the benefits to that person's sexual partners -- and to society, as a whole -- it seems like sort of a slam dunk.

Myles Helfand: Dr. Sax, anything you want to add to that?

Paul Sax: I'd just like to reinforce the last point that Joel said, and mention that it's had an interesting effect in the clinic: Even though it's a public health benefit, this prevention transmission, it turns out to be very motivating for your patients, also.

It's really unusual these days to have someone, once they hear that information, say, "Still, treatment's not for me." Because I guess, deep down, people understandably don't want to be contagious to others. They want to be less contagious. And so 052 -- even though it's kind of new, and we knew it was going to work -- still, having those results out there really changed things a lot.

One other factor is that there's a long-standing belief that, if you use your treatment now, when you're healthy, you're not going to have anything left when you're sick -- because of the idea that eventually everybody fails. But it really turns out that's not the case.

We now know that, over the long-term, people who are successfully treated can remain successfully treated, probably indefinitely, provided they continue their medications. The development of multi-class resistance is a thing of the past, and more linked to the inadequate therapies we had in the '90s than anything about treating HIV.

Joel Gallant: Yeah. In fact, there are some studies now that are showing that people who start with higher CD4 counts are less likely to develop resistance, and less likely to develop toxicity, than people who start with lower CD4 counts. So that just emphasizes the point that Paul made.

Myles Helfand: To what extent is that related to an increase in the potency of the regimens that we have, versus, say, increased levels of adherence, thanks in part to the proliferation of single-pill, once-daily regimens?

Joel Gallant: I think it's both. I think the regimens clearly are more potent; but they're also much easier to adhere to, because of their simplicity and greater tolerability.

In the past, people could do well with combinations of indinavir [Crixivan] and stavudine [d4T, Zerit] and lamivudine [3TC, Epivir], but it was really challenging. And the toxicity was really significant. So even the most adherent patients were going to have problems with those regimens.

Now, we're typically in clinical trials seeing efficacy rates of 90% or above, which is just unprecedented, compared to earlier regimens. It's a combination of the better drugs and easier adherence.

Paul Sax: The other factor is that the use of boosted protease inhibitors [PIs] had almost eliminated the whole issue of new onset protease inhibitor resistance. It's really remarkable, when you think about your clinic population and who the patients are in your clinic who have triple-class failure -- the old-fashioned, triple-class failure [i.e., to PIs, NRTIs and NNRTIs] -- and have high-level PI resistance. Really, none of them started treatment after 2001 or so, when lopinavir/ritonavir [Kaletra] became the standard PI.

So it's really uncommon to see any new cases like that. They do happen, but they're not happening frequently.

The U.S. vs. Great Britain (and the World)

Myles Helfand: You mentioned HPTN 052 and this increasing awareness that "treatment as prevention" is a real, actual thing -- it's a viable strategy for helping prevent the further spread of HIV. The British HIV Association relatively recently updated its own treatment guidelines, and specifically avoided a recommendation that people start HIV treatment at greater than 500 CD4+ cells.

Dr. Sax, why are the U.S. guidelines more aggressive than the British guidelines in that sense?

Paul Sax: I think it's a matter of degree. First of all, there is unanimity among HIV specialists that, if you have limited resources for paying or covering antiretroviral therapy, you get much more benefit per person if you delay treatment. That's excluding the transmission piece. So if that's the case, then actually recommending therapy only for people with lower CD4s makes sense.


But if you look at it differently -- if you say, "Well, we do have the resources to pay for these treatments (and I would argue we do), then there is benefit accrued along the entire CD4 spectrum. The one that's most obvious are people with AIDS; but then, of course, there are these other benefits at higher CD4s that Joel has alluded to.

So I think it has a lot to do with resources. We have problems with our health care system in the United States. No one needs to go through them. But one of the problems we don't have -- at least right now, knock on wood -- is paying for HIV treatment among those who are in care.

Myles Helfand: I don't know if the people on ADAP [AIDS Drug Assistance Program] waiting lists would necessarily agree with that, though.

Paul Sax: Well, let's take a look at that. I think it's reasonable to drill down and say: How many people in the United States who are in care are actually denied HIV therapy?

And the number is actually very small, even in the worst-case states -- you know, the southeastern United States. I would argue it's much lower than people who are denied care for other health care conditions than it is for HIV. So, by contrast, if you talk about some of the very expensive biologics to treat rheumatoid arthritis, or you're talking about neurologic conditions that have very expensive treatments, that's much less likely to be covered than HIV therapy.

One reason why we've been so good at having a system for treating HIV is because treatment is so darn effective. It's very hard to argue against it.

So, yes: There are individual cases and individual states where there are waiting lists. But those are relatively small numbers, given the size of this country.

Joel Gallant: One of the issues is that some of the benefits that we accrue by treating very early may be benefits that you wouldn't notice in the short term. It's not that you're going to prevent people from developing pneumocystis, but that you may be lowering a long-term risk of things like cardiovascular disease or bone disease. We may not see those benefits for several decades. And, of course, when people who control the purse strings make decisions, they're not really interested in those long-term benefits.

But there's also the financial benefit of preventing transmission, since it's obviously much cheaper to prevent a new case of HIV infection than to treat it for life. Paul has more experience with cost effectiveness analyses than I do, but there have certainly been people who have looked at the cost effectiveness of treating to prevent HIV and it's generally come out to be very favorable.

Myles Helfand: Yet then we look at WHO [World Health Organization] guidelines, which are about as conservative, relatively speaking, as you can get when it comes to HIV treatment guidelines globally. There's been a lot of discussion over this, seen over the past few years, regarding the nature of WHO guidelines, and to what extent they should be capitulating to public health and policy realities, versus just saying, "This is what people need."

You had mentioned just a little bit earlier, Dr. Sax, that if we have the resources to be able to pay for treatment that we should do it. Why wouldn't treatment guidelines be irrespective of the ability to pay?

Paul Sax: Well, that's a very interesting debate -- whether guidelines should include issues of cost and resource allocation. I'm of the opinion that introducing that topic is important.

But we do live in that finite world, as far as resources are concerned -- don't get me wrong. For example, the use of tesamorelin [Egrifta] for visceral adiposity is not generally covered in AIDS Drug Assistance Programs. But at least here we have a fairly good system, as fragile as it might be, to pay for treatment for everyone.

If you don't have that -- and internationally they don't, and they have to prioritize -- you really do get much more bang for the buck by treating people with more advanced disease.

Interestingly, there's this paradox: In resource-limited settings there are more aggressive opportunistic infections that occur at higher CD4 cell counts, [and thus] the benefit, as far as preventing AIDS complications, is greater in those settings than it is in the United States. Which is one reason why, even after HPTN 052 showed a clinical benefit for people with CD4s greater than 350, some people have said, "Well, that's only because those are mostly cases of tuberculosis. How can you show that same benefit in Western Europe and in the United States, where there's much less TB?" And that is a fair argument.

Joel Gallant: On the other hand, there's also greater risk in toxicity to early treatment in a developing country setting than there is here. For example, in most of the countries that the WHO guidelines are written for, typical starting regimens involve d4T or AZT [zidovudine, Retrovir], which we don't use much. If they're monitoring viral load at all, it's very infrequently, so patients often have pretty extensive resistance by the time they've moved to second-line therapy.

So, in that setting, under those conditions, it probably wouldn't make sense to treat at very high CD4 counts. Because you're putting patients at great risk for relatively small benefit.

In contrast, here in the U.S., where the regimens are very safe and we have good monitoring, there is a benefit that outweighs the risk.

Paul Sax: Yeah. There's a really nice paper that was just published in AIDS, looking at the response to ART [antiretroviral therapy] in sub-Saharan Africa, and looking at how a CD4 greater than 500 appears to confer an even greater benefit there than it does in the U.S.

When Not to Start Therapy

Myles Helfand: OK. So, here in the U.S., the guidelines now unanimously recommend that therapy be initiated at any CD4+ cell count, which is a pretty ubiquitous recommendation. But what are the exceptions to that rule?

Are there cases in which a clinician, on a case-by-case basis, should be second-guessing that? Or should they think, "As soon as somebody comes into my clinic" -- or my hospital, or my practice -- "and is either diagnosed with HIV or just presents with HIV, I need to start them on treatment"?

Joel Gallant: Well, remember, it's not a rule, it's a guideline, a recommendation. And I think one of the most obvious exceptions would be somebody who is simply not ready to start -- who either doesn't want to start or who is going to have problems adhering to therapy.

In patients with very high CD4 counts, this is never an emergency. So you have time to prepare the patient and to make sure that everything's set up to maximize adherence. We shouldn't look at this as something that has to be done on the first visit. Patients need to buy into this, and be ready to do it.

The other exception for me would be the elite controller who has a high CD4 count -- the person whose viral load is already undetectable and the CD4 is stable and normal. It's possible there could be some benefit to treatment for those people, but it sure is going to be hard to prove it.

Myles Helfand: How can a clinician know whether a person is going to be an elite controller? Is it necessarily an indication if they present with a high CD4+ cell count -- say, above 600 -- and an undetectable, or very low, viral load? That doesn't necessarily mean they are an elite controller, right?

Paul Sax: No. Absolutely right. We all have had experiences -- in particular, people with recently acquired HIV, where they start out with very favorable-looking prognostic numbers, both by clinical experience and cohort analyses, and they get virologic escape from the immune control. That can occur over time. There's no easy way to tell whether someone's going to end up having that phenotype permanently.

There are some correlates, like the HLA-B*5701 -- ironically, the same one that goes along with abacavir [Ziagen hypersensitivity] -- and the CCR5 heterozygotes. But nothing absolute.

Joel Gallant: Of course, you don't need to be able to predict the future. You can simply say, "I'll treat when there's either CD4 progression or viral escape."

Myles Helfand: Well, how do you balance? Is there a checklist that a clinician can go down, where he or she begins to say, "The CD4+ cell count is 350; now I begin pressuring the patient and talk to the case manager about getting the patient more effective support"? How do you begin to make that differential between "OK, this person is fine; I don't need to push too much" and "This person really needs to get into treatment"? Dr. Sax, let's go with you.

Paul Sax: This is going to sound perhaps overly obvious. But there are people that you really have to twist their arms, because you have to remind them that even though they feel well, HIV is still potentially a fatal condition, and AIDS is a fatal condition; and that, in order to avoid that, they'd have to go on antiviral therapy.

Most of the time, you can make the case very persuasively and get people to buy into treatment. I like the way Joel phrased it before: It's for those people where the urgency of starting is considerably greater.

You can either use, if they're very scientifically oriented, all the science; or you could use pictures. You could use all kinds of analogies, like saying, "Look. This is like treatment of very severe high blood pressure. You may feel totally fine, but what we're trying to prevent is a devastating stroke." And getting people on treatment, most of the time, works.

I will say that there is a subset of people who simply won't take antivirals. Based on cohort studies, based on our own clinic, and based on Ryan White reports, they're probably 10% of most practices. A lot of them don't show up [for their appointments] -- or, when they do show up, they are sick. It's very hard.

I think one reason it's very hard to manage them is because the causes of that kind of behavior are so diverse -- not just one thing. It's usually multiple things.

Joel Gallant: I think also, patients sometimes come in and are surprised when we recommend treatment. Or I hear other clinicians say, "Well, my patients, when I tell them they should start treatment, they balk at that."

Well, of course! They've been taught the lessons that we've taught them over the years: That you don't need treatment till you're sick, or you don't need treatment till your CD4 falls. When we tell them we're now recommending treatment for everyone, it comes as a surprise.

But we can easily re-teach them. They have an infectious disease; we have a treatment for it. It doesn't take a big leap to understand that you should be treating this infectious disease. It's just a matter of undoing the lessons that we've been teaching them over the years. And I think that can be done.

HIV Treatment Is Only Part of the Solution

Myles Helfand: This is probably also an important reminder for all of us that connecting a patient to relevant services, support services and education services, is so critical in HIV. In many cases, it's more than just going to your HIV physician every three to six months and getting a viral load test and talking about your health. It is also being able to be connected to the level of care that can get you educated more about HIV and its treatment, that can maybe get you some help for depression if you're experiencing it, to get help with unstable housing if you're experiencing it, to get help with just taking care of your loved ones and your family if you need it. The kinds of things that can be a barrier to starting therapy in the first place.

Paul Sax: Absolutely. One thing that's interesting: We were doing a quality assurance program, and we got a look at the proportion of patients here who are prescribed ART. I can't remember whether it was 93% or 91%; most of the patients in our clinic are either on therapy, or prescribed therapy. And that actually was reflected in a national study that was presented at CROI this year. So, among people in care, we're already very close to treating everybody.

Another way we know that anecdotally is because the basic science researchers we collaborate with are always looking for people not on treatment to help their pathogenesis studies along; they're looking for specimens. And it's getting harder and harder for them to find them. I'm sure this situation is similar at Hopkins.

Joel Gallant: Yeah, exactly. If you look at the CDC [U.S. Centers for Disease Control and Prevention] studies that have shown anywhere from 19 to 28% of Americans with HIV have undetectable viral loads, the problem is not that they are in care and not getting drugs, or that they're on drugs that aren't working; the problem is, they're not in care, or they're not diagnosed. That's where our gap is.

Once people are in care, they tend to get put on treatment and they tend to suppress their viral load.

Paul Sax: Exactly.


Revisiting Treatment Interruption

Myles Helfand: Speaking of getting on treatment and staying on treatment: Structured treatment interruption used to be a fun topic of much discussion and debate. Earlier in the last decade it fell apart, due largely to the findings of the SMART study, but also due to the increased potency and decreased toxicity of existing treatment, making it less likely that a person would want to stop.

There have been a few studies that, over the last year or two, have started to explore in more detail the idea of starting treatment during acute infection. These are people who start treatment immediately in order to achieve a level of viral suppression, maybe a stabilization of CD4+ cell count. Then after, say, six months or a year, they stop treatment for at least a few years to reduce the risk of long-term toxicity.

How do either of you feel about that approach? Is there any merit to it? Or do we just not know enough yet to be able to recommend that as an actual approach to treatment?

Joel Gallant: I'm not too excited about it. There is some evidence that early therapy, right after acute infection, may delay your need for antiretrovirals later. But we were doing a study like this, and when the time came for people to stop, they were often doing extremely well. They had high CD4s; they had undetectable viral loads; they were having no side effects; and the idea of stopping made no sense to either the patient or the clinician.

I think that treatment interruption studies are a remnant of the days when all we cared about was keeping your CD4 above a certain threshold, and nothing else mattered. Ever since the SMART study, we've realized that there's more going on than just CD4; that rebound viremia is not good for you.

So, for me: I just don't see the point of interrupting therapy that's working.

Paul Sax: Yeah. Those studies that you refer to, Myles, are interesting: Two randomized clinical trials for acute infection, where you had strategies in interrupting therapy.

I would be very surprised if any such study could be started today, simply because interruptions have consistently been associated with more harm than good. And there's no reason to really think why acute infection would be necessarily different from chronic infection, in that regard.

So even though, yes, there was benefit from treatment, as compared to no treatment, I bet you the benefit would have been even better if they had stayed on treatment. Believe me, and getting back to what you were saying before, we understand that these treatments, even though they're safer, they do have side effects. One of the big challenges in HIV medicine is managing the long-term side effects of these treatments, and knowing when is the proper time to consider a change, even with virologic suppression. But that's a whole different topic.

Myles Helfand: Yeah. That's a topic for another day, unfortunately.

What the Future Holds for "When to Start"

So, where do we go from here? We used to refer to the debate over starting treatment as a pendulum that, for the better part of a decade, had felt like it swung in one direction against the initiation of therapy, due to toxicity issues and resistance concerns. And now, we are about as far in the direction of, "Yes, go ahead and start treatment," as we have ever been.

Are we done swinging back and forth? Looking into the crystal ball, where do you guys think you might be headed, in terms of when to start? Dr. Gallant?

Joel Gallant: I think the pendulum's done swinging. But then, you always think the pendulum's done swinging. You always think that what you know now is what's correct. And then you can be surprised sometimes.

The one thing that's out there that we don't know about yet is the START study, which is a randomized, controlled trial of when to start -- of earlier versus somewhat delayed therapy. Many people really objected to the guidelines moving in the direction that they did, without the data from the START study. They said that this would discourage participation in the START study.

While that may be true, I think the guidelines needed to do exactly what they did. Because the guidelines panelists felt that the information was sufficient to make these recommendations. And it would have been unethical to withhold that recommendation just for the sake of one study.

What happens if the START study shows a benefit to early treatment? Well, that will just confirm what we've already done.

What happens if there's no difference between early and delayed therapy? I don't think that will change people's minds. Because many will say, "Well, the benefits are long-term benefits. You're not going to see them in a trial, where your follow-up is, you know, six years." And in addition, it's not going to be measuring transmission.

It seems unlikely that the START study is going to show worse outcome with early therapy. If that happened, that would be a game changer. [But] I'm pretty comfortable that we're going to stay where we are.

Paul Sax: One issue that has come up with the START study is the question of how representative it is of patients in practice. That's because, when you look at the subset of people who are both (a) willing to be randomized and (b) cared for by providers who are going to randomize them or refer them for randomization, it may not be similar to the overall group who's out there.

It will answer some questions, no doubt. I think it will be fascinating to see what it shows. But it may not be the last word on this.

The other thing is, getting back to this cost issue: One thing that will be happening over the next few years is increasing availability of generic HIV medicines [in the U.S.]. Just like treatment of hypercholesterolemia becomes less expensive when statins go off-patent, the same will happen with HIV therapies.

Myles Helfand: At the same time, though, we have new therapies that are going to be approved. As of this recording, we've got the "quad" drug that's right on the horizon that is going to have a pretty long patent life, even as existing first-line meds -- like efavirenz [Sustiva, Stocrin], for instance -- come off their patent within the next 12 to 24 months, probably. To what extent do new drugs, like the quad, change the conversation at all?

Paul Sax: The safer and more effective the initial intervention, the easier it is to start treatment. Clearly, we need longer-term data on the quad and more clinical experience after approval, assuming it is approved, to say that that's the answer.

But for some people, they don't want to take a medication that has any CNS [central nervous system] side effects, like [those sometimes caused by] efavirenz. Or they may have a viral load that's too high to go on a rilpivirine [Edurant]-based treatment.

There's no perfect treatment out there. Obviously, if the medications were perfect, then everyone would be on treatment. It's moving the field forward incrementally, giving these new options.

Closing Thoughts

Myles Helfand: All right. Do either of you have any closing thoughts? Anything on the when-to-start question that we haven't touched on, or that you'd like to say before we close up? Dr. Sax?

Paul Sax: I think it's important to continue to have these dialogues about this issue. I know that I've heard from some of my European colleagues that they don't find the data as compelling as we do. If we keep an open mind, we're more likely to learn. So even though Joel and I are in agreement on this, I have to allow that this [starting HIV treatment early] is not one of those obvious situations, like it is when you're treating people with more advanced immunodeficiency.

Joel Gallant: Yeah. I think the Europeans, and the British physicians and guidelines panelists, very much have to weigh cost in their decisions. In systems where there's a fixed pot of money and they have to decide the greatest good for the greatest number, it's impossible to ignore cost.


In the U.S., we've had the luxury of not having to think too much about that. We've been able to be purists, and just talk about what we think the scientific data show, in terms of our recommendations. Maybe we've been deceiving ourselves that way.

But my own opinion is that, when you consider both the treatment and the prevention aspects and benefits of antiretroviral therapy, we can't afford not to treat people. And I do think we have the money to do it.

Paul Sax: Exactly. The problems with the U.S. health care system are so huge that I don't even want to start. But the reasons why it's so expensive do not have to do with antiretroviral therapy. I think that's pretty much all I need to say.

Joel Gallant: Absolutely. I mean, antiretroviral therapy may be pricey, but it has been shown to be one of the most cost-effective medical interventions that we have. And that's without the prevention benefit; that's just looking at treatment. If you add in the prevention benefit, it just seems overwhelming.

Myles Helfand: On that sobering but slightly optimistic note, we'll wrap up. Thank you, Dr. Gallant and Dr. Sax, for taking part in this discussion.

This transcript has been edited for clarity.

Myles Helfand is the editorial director of and Follow Myles on Twitter: @MylesatTheBody.

Copyright © 2012 Remedy Health Media, LLC. All rights reserved.

This article was provided by It is a part of the publication HIV Management in Depth. You can find this article online by typing this address into your Web browser:

General Disclaimer: The Body PRO is designed for educational purposes only and is not engaged in rendering medical advice or professional services. The information provided through The Body PRO should not be used for diagnosing or treating a health problem or a disease. It is not a substitute for professional care. If you have or suspect you may have a health problem, consult your health care provider.