Metabolic Complications of HIV Treatment

Translating HIV Research Into Practice

Table of Contents


As our patients are surviving much longer with their HIV infection, we've been seeing a troubling series of body shape and metabolic abnormalities appear in these patients. They're called HIV lipodystrophy, for really lack of a better name.

Patients who have the HIV-associated lipodystrophy may have one or all of the following changes:

  • Deposition of visceral or central fat, and the wasting or atrophy of the subcutaneous fat; that's the fat between the skin and the muscle.
  • Metabolic abnormalities -- the atherogenic lipid profile that is seen most often is a low HDL cholesterol and a high level of triglycerides. In patients who are treated we may also see elevations in total cholesterol and LDL, but low HDL and high triglycerides are the most common.
  • While we don't see frank diabetes generally, we do see either insulin resistance or glucose tolerance.

These changes do not have to occur in concert. An individual may have one or more of these, and the data we have to date suggests that the body shape changes clearly are not linked either epidemiologically or etiologically, but the changes in body shape may be more associated with some of the metabolic abnormalities.

Click to enlarge.
Click slide to enlarge

The metabolic syndrome is an occurrence of metabolic abnormalities that has a strong independent risk for cardiovascular disease, and many of the components of the HIV lipodystrophy syndrome appear to present much as the metabolic syndrome, which in the HIV-negative population (the general population) has also been called "Syndrome X."

Metabolic syndrome is defined, whether in the HIV population or not, by the presence of three of the following five abnormalities:

  • A low HDL, and this is a different abnormal level for men and women. In men, the abnormal level is less than 40 mg/dL and in women it's less than 50 mg.
  • A high triglyceride level, which is the same for either gender.
  • An elevated waist measure, which does differ for men and women. In men, it's abnormal when the waist measure is greater than 102 cm; in women when it's greater than 88 cm.
  • Elevations in blood pressure, which are defined as a blood pressure greater than 130/85.
  • An abnormal glucose metabolism, which is defined as a fasting glucose of greater than 100 mg/dL, or being treated for diabetes mellitus.

The metabolic syndrome is a risk factor for cardiovascular disease, and the medical outcome of most concern for HIV-infected patients with body shape changes, or metabolic changes, is an increased risk for cardiovascular disease. The data we have today that suggests that this risk is real include an increased number of cardiovascular events in the D:A:D study. There have been a number of studies that have defined abnormal surrogate markers for cardiovascular disease in HIV-infected individuals, which presumably defines an increased risk for cardiovascular events at some point in the future. As an aside perhaps, there are a number of studies that suggest that HIV is associated with an elevation in high-sensitivity C-reactive protein (hs-CRP), and C-reactive protein is also an independent risk factor for cardiovascular disease. We must also acknowledge that as our population ages, they are going to have the potential for increased cardiovascular disease. Because of improvements in therapy, our population is surviving longer and is aging.

The risks for cardiovascular disease in an HIV-infected individual can be divided into three large categories: the host, the viral factors and treatment factors. The host factors are either modifiable or unmodifiable, and things like genetics and age are obviously not modifiable risk factors. Genetics -- we don't really have access to tests clinically, but the most reliable way for us to try to determine whether there is a genetic risk in a patient is by determining family history. Many of our HIV-infected patients also have modifiable risk factors, such as smoking. They may have poor dietary quality and they may not be exercising as much as we would like. In terms of HIV as a risk factor itself, we've just discussed the potential for increased C-reactive protein to be contributing to risk. We think of C-reactive protein almost as a surrogate marker of HIV disease burden over time. As with the changes induced in CD4 count and viral load by therapy, we don't have the benefit of using CD4 and viral load as a marker of overall disease burden.

Other people have used the nadir CD4 count (the lowest that a CD4 has ever gotten) or the height of viral load at any point in the past as additional surrogates for overall disease burden. These are the best ways we have at present of really trying to understand how much of an impact the HIV disease has played in an individual patient. As we just discussed, there are low HDL levels and high triglycerides in the untreated HIV-infected patient that may worsen with therapy. But HIV itself may contribute to cardiovascular risk by inducing these lipid abnormalities on its own. In terms of treatment, we know that a number of pieces of data suggest that selected protease inhibitors lead to even more atherogenic lipid profiles with further increases in triglycerides, further decreases in HDL cholesterol and increases in total cholesterol, and that select antiretroviral agents from a number of different classes may contribute to insulin resistance.

We know that the D:A:D study suggests that there is an increased risk for cardiovascular events with any protease inhibitor therapy. There certainly are a number of epidemiological reports that suggest that some of the body shape abnormalities (such as an increased waist measure) that may put an individual at risk for cardiovascular disease may be associated with treatment as well.

Case Study

The case that we're going to discuss this evening is a 37-year-old male infected with HIV. He is not an injection drug user. He is treatment naive as he has a CD4 count of 419 and a viral load of 53,867. His fasting triglycerides are elevated at 567 mg/dL (that's greater than the 150 cutoff for normal). His total cholesterol is just slightly above normal. He should have a cholesterol of about 200, and his is 203. His HDL cholesterol is markedly reduced at 14 mg/dL. His LDL cholesterol at 76 is well within the normal range, as is his fasting glucose at 83. The patient does acknowledge that he has smoked a pack a day for the last 20 years.

When we begin to look at his nutritional or body composition parameters, his body mass index at 23.9 is well within the normal range. For people who are not used to or familiar with the body mass index as a measure, body mass index is used as a way to adjust weight for height. A normal BMI is considered to be between 23 and 27, so his is nicely in the normal range. Greater than 27 is considered to be overweight and greater than 30 is considered to be obese. Our patient's waist measure is 85.3 cm. We use DXA in this patient to determine total body fat; at 17%, that's well within the normal range for a man of his age. His body fat should be between 15 and 19%. His DXA total fat came out to be about 13.5 kg, and predominantly that was in his trunk. He had 7.8 kg of fat in his trunk as determined by DXA. His C-reactive protein was elevated at 2.78. This is very close to the cardiac risk value of 3, and his blood pressure was in the normal range at 107/63.

Click to enlarge.
Click slide to enlarge

The next slide is just an example of a DXA scan showing how we use it to determine body composition. We're all much more familiar with using DXA for determining bone density, but in this case with the modification of software, the very same scans can be read for determination of lean body mass and for fat mass. What DXA can't do is detect the difference between fat that is under the skin and fat that's within the abdominal cavity. We get a measure of trunk fat, but we can't say where that fat occurs in the trunk. And you can see the lines drawn on these two scans, so that you can see how we determine appendicular fat and truncal fat, really just by calculating the lean [body mass] and fat in the areas sort of indicated by the white lines on the scan.

In terms of considering what kinds of interventions we'd want to make for our patient that's just been presented; his HIV does not meet treatment guidelines. His CD4 is too high, and he is totally asymptomatic. He does not meet the definition of metabolic syndrome, as he has only two abnormalities in his laboratories, and those are his triglyceride level and his HDL level, which, again, are the most common lipid abnormalities that are seen in HIV. His weight and BMI are well within the normal range. His waist measure is in the normal range. His body fat is in the normal range. With all of that said, it's his total cholesterol by a hair, his triglycerides and his HDL that are abnormal, and these may increase his risk for cardiovascular disease, and his CRP is high and it is approaching the cardiovascular risk level, which is any CRP level greater than 3.0. The concern is obviously, with these abnormalities in a very young individual, that there may be a potential risk for cardiovascular disease in the future.

If we want to try to determine an intervention for his metabolic abnormalities, our goal would probably be to try to decrease his triglyceride level and increase his HDL cholesterol. We can do this in a number of different ways that are summarized on this slide. We could certainly discuss the potential for trying to improve his diet quality and increase his exercise, which are clearly effective in decreasing triglycerides and increasing HDL in the non-HIV population and have been shown to be effective in the HIV population. However, they may be more effective in individuals who are overweight, and this individual is not overweight. This individual is eligible to be considered for fibrates, as he has no antiretroviral therapy on board. He is not at risk for drug/drug interactions with the cytochrome P450 system, which both the fibrates and many of the antiretrovirals use. He could be considered for niacin therapy, but the side effect profile of niacin, even the slow release, is often severe enough that patients don't wish to continue this therapy. He is a candidate for fish oil. The intention with the use of fish oil is to increase the intake of omega-3 fatty acids and thus, change the ratio of omega-3 to omega-6 fatty acids that are in the diet with an attempt to (by improving the omega-3 ratio) change the lipid metabolism. A more novel or potentially experimental approach would be to think about treating his HIV, even though his CD4 count doesn't warrant it, in an attempt to decrease inflammation since he does have the high CRP. But that's not suggested by the guidelines or necessarily consistent with standard of care.

We elected to treat this patient with omega-3 fatty acids. He received 4 grams of omega-3 in fish oil every day for six months. It is possible to give fish oil by different routes. Patients can institute dramatic alterations in their diet. These alterations require the ingestion of fish at one or two meals a day every day and can be quite unpleasant to some of the patients who are not that fond of eating fish. It's possible to give omega-3 by fish oil capsules, and there is also a purified prescription omega-3 that doesn't have the rest of the fish oil. With any of these routes of administration, there has to be some concern for potential contamination by mercury as much of the omega-3, or the fish oils, are obtained from salmon. There is a potential for mercury contamination even in the farmed salmon. We went with the fish oil capsules for our patient, and we also used our very experienced HIV nutritionist to counsel this patient on changing his diet, where he was encouraged to eat salmon two to three times a week. He was encouraged to include walnuts in his snacking regimen, and he was encouraged to increase his fiber intake. Any time you try to do a dietary manipulation like this in a patient, it is important to have an experienced nutritionist onboard so that the diet can remain appropriate in terms of total caloric intake, as well as percentages of dietary fat.

As an aside, we encouraged this gentleman to stop smoking, thinking that this intervention would be a good time to start talking about modifiable risk factors as well.

Click to enlarge.
Click slide to enlarge

At six months what we found in this patient was he still remained reasonably stable in terms of his HIV. His CD4 count is 391, which was a change of only 28 cells. His viral load was nearly identical at 53,406, really not a change at all. His fasting triglycerides, however, had decreased to 245 mg/dL, which is a change of 322 mg/dL. His total cholesterol had gone down to 192 mg/dL, a change of 11 mg/dL, and his HDL cholesterol had risen to 23 mg/dL, an increase of 9 mg/dL. His LDL cholesterol had also gone up. His fasting glucose was unchanged, and we did not have an impact on his smoking, even though we referred him to a smoking cessation program and encouraged him to attend. He elected not to do that.

In terms of his body weight and composition, his BMI and weight were incredibly stable, really didn't change at all. His waist interestingly decreased 4.5 cm to 80.8. His DXA [bone density scanning, also called dual-energy x-ray absorptiometry ] total body fat went down nearly half a percent, from 17 to 16.6, so a mild change. His DXA total body fat decreased about 2.5 kg, which suggests that his body fat percentage may not have changed dramatically, but he may have put on some lean [body mass]. His DXA trunk fat had decreased about a half kilogram; his waist measure was also decreasing.

Interestingly, his CRP was 1.66 after the six months of therapy, which was a change of 1.22, and this is presumably through the omega-3 changes, arachidonic acid pathway metabolism and decreases the production of arachidonic acid, which is an inflammatory mediator. His blood pressure remained well within the normal range after his fish oil therapy.

In summary, we found that in this patient 4 grams of omega-3 fatty acids daily for six months without any change in weight really for the patient, so not related to a weight reduction diet, led to a successful 57% reduction in triglycerides, although I point out that this is still not within the normal limits for triglycerides. He had a successful increase of 64% in his HDL, although this was also still not within the normal limits. His total cholesterol decreased 5% and was reduced to the optimal level. His LDL cholesterol increased, although it remained near the optimal level. He had a substantial decrease in his waist circumference and his truncal fat, and also successfully decreased his CRP.

It is possible to successfully improve these metabolic parameters in HIV-infected individuals. I think we get discouraged sometimes because these triglycerides and HDL [abnormalities] are so difficult and so persistent. We assume that this patient had either genetic or HIV-related metabolic disorders, as again he was treatment naive. He was not on any antiretroviral therapy. We point out that this is a non-drug intervention that led to a significant improvement in his HDL, triglycerides and total cholesterol, as well as the decreases in his trunk fat, total fat and C-reactive protein. But this patient obviously is going to need to stay on his fish oil in an effort to continue to reduce his cardiovascular risk. We still need to try to encourage him to stop smoking, and we need to follow up his cholesterol, his triglycerides and his HDL in an attempt to determine if we can get these even further more towards the normal range.

I would just like to mention that we have seen similar responses to fish oil in treated HIV-infected individuals with high triglycerides and low HDL. The effect can be reproduced in people who have treatment-induced alterations in their lipid profile. A number of studies again suggest that it's the high triglycerides and low HDL that are the most common lipid abnormalities in HIV-infected individuals, whether they are treated or untreated, and that these are the least likely to be treated. The use of statins is the most common lipid intervention for HIV-infected individuals, and these are much less likely to have an impact on triglycerides and HDL than they are in total cholesterol. So that we can expect to see improvements in these metabolic parameters with fish oil, again stressing that this is a non-pharmacologic intervention that shouldn't lead to any drug/drug interactions.

Click to enlarge.
Click slide to enlarge

When this patient needs to begin antiretroviral therapy with his history of elevations in triglycerides and decrease in HDL cholesterol, we need to try to construct a metabolically neutral antiretroviral regimen. There are a number of possibilities with the drugs that we have available to us now to treat HIV. He would probably warrant the use of dual nucleosides using those that are more metabolically neutral or metabolically beneficial, and either an NNRTI, which may also have some benefits in individuals who have low HDLs and may lead to an increased HDL.

We could consider using dual nucleosides with atazanavir, although atazanavir is not likely to improve further lipids that were abnormal off therapy; it is most remarkably able to improve lipids that become abnormal on therapy. But atazanavir should not lead to any worsening of his lipid profile. If at all possible, we would avoid the use of ritonavir, even at boosting doses, for as long as we could, to avoid the exacerbation that ritonavir might be expected to produce on his triglyceride level and his HDL abnormalities. There is not sufficient data to date to be able to say anything about the expected impact of either integrase inhibitors like raltegravir, or CCR5 inhibitors like maraviroc, and what impact they might be expected to have on these lipid parameters over time.

Questions & Answers

That is where I wanted to leave the discussion. I have some questions here that I'm happy to answer and then people can ask any additional questions.

The first question is: At what stage in the course of disease do the effects of HIV-associated lipodystrophy and metabolic syndrome typically first appear?

I think that that seems to be genetically regulated, because different patients may have metabolic changes very early on; others will not have metabolic changes until they've apparently been infected for a much longer period of time. We have noticed that these are likely to occur more frequently in treated individuals, so there is a very strong presumption that treatment plays a role in the induction of these syndromes, although it's possible that the treatment just allows individuals to survive long enough that we see these. So there is not really a time frame that we can pin on the expected appearance of this syndrome.

The second question is: Have there been any large randomized clinical trials showing the efficacy of fish oil in people with high triglycerides?

The answer to that is yes. There are studies in both HIV-negative and HIV-positive individuals that suggest that fish oil is quite successful in reducing triglycerides in individuals.

Why 4 grams of fish oil a day? How was this dosing arrived at?

That's the dose that many of the studies have used. If you don't use enough fish oil, you don't begin to alter the lipid metabolism because you don't change the omega-3/omega-6 intake ratio, and you really have to shift that ratio towards the omega-3 fatty acids in order to begin to see the beneficial effects. There have been essentially dose-response studies done that suggest that 4 grams of fish oil is the dose that we really need.

If the patient is not on HIV therapy, why not give him statin? Aren't they generally more effective?

Statins would be much more effective [at lowering] total cholesterol, [but] his total cholesterol was actually not a problem. It was his HDL cholesterol and his triglyceride level. And if we can avoid pharmacologic therapy since he will have to be treated and keep his medication burden down, we prefer to do it that way. Certainly using pharmacologic interventions, either a fibrate or niacin, in this patient would be a viable way to go.

Can I provide an interpretation of the SMART study finding of increased cardiac events in patients who stop their antiretroviral therapy?

I think it's a very important study. I think we don't have enough information to really say conclusively what that means, but it certainly says to me that the uncontrolled replication of the HIV virus does have the potential to induce cardiac events. It's almost something over and above these long-term atherosclerotic changes that might be seen in HIV patients. It's almost like these patients are having acute coronary syndrome because the time frame was relatively short. I think we're going to learn more in the future about the role of the virus, as well as the role of therapy and contributing to cardiovascular risk.

Another question is: What would be the recommendation for an HIV-infected patient who is exercising seriously and has a good quality diet but still has been dealing with fat accumulation in the belly area?

This is a very tough question. This is a condition that a number of our HIV-infected patients have been dealing with. If at all possible, we try to get these individuals into clinical studies that are using growth hormone releasing hormone. We've been part of metformin trials. We've done growth hormone studies. We try to get people into studies that propose different interventions for this visceral fat accumulation. For people who say that they're exercising and have a good quality diet, we still really like to have them meet with an exercise trainer or a specialist because people may have differing ideas about how much intensity needs to be in an exercise regimen. We need to make sure that there is a good balance of both aerobic exercise and progressive resistance training because it's a combination of the two that seems to have the most impact in HIV and this visceral fat accumulation. There are a number of controlled trials that suggest that it's quite beneficial. And a patient's interpretation of what a good diet is and an experienced nutritionist's interpretation of a good diet may also vary significantly. We found that even patients who are quite sure they're eating a good healthy diet may benefit from meeting with a good nutritionist.

We've got a lot of questions here.

There is some research that suggests omega-3 capsules are not as effective as omega-3 directly from fish or nuts, etc. Given that the patient has had good results while taking omega-3, how effective do you think the capsules are in comparison to omega-3 directly from food?

I can't really comment on that. I think it's the amount; we did encourage this gentleman to eat fish and to take walnuts as well, and he was reporting that he was doing that, so he may have been boosting his omega-3 intake even higher. But I don't believe there has really been a head-to-head comparison of fish oil capsules with foods. And again, a number of patients for whom we've tried to just induce the omega-3 load with diet, have found the diet very difficult to maintain. It also can be very expensive for patients.

Another question: I have a patient who has lipoatrophy as a result of d4T use, and he is now on Atripla, but we've seen no reversal of his lipoatrophy. Any suggestions?

This is a very tough one. The studies that have looked at this and have done quantification of fat mass as people come off of d4T suggest that it may take up to two years to begin to see even a mild degree of reversal of this lipoatrophy. It may just be more time. The other thing that appears to happen with d4T use is that it can induce some insulin resistance, and fat atrophy appears to be associated with insulin resistance. It would be interesting to just check this patient's insulin sensitivity, do an oral glucose tolerance test and see if he/she is insulin resistant, because that might present some potential pathways for intervention.

Are there more intensive strategies that have the potential to normalize triglycerides and HDL?

Certainly you can use many of the potential interventions we discussed concurrently and try to improve these even more. There was an ACTG study that suggested that if you treated with a fibrate and didn't see a complete response and you added fish oil to that, you could intensify the results. But to use these agents in concert, you should be able to continue to get improvement, although I'm still not sure that you can necessarily get every single patient back to normal limits. You might be able to get them much closer.

With the change in guidelines, would I treat this patient?

That's a good question, and I think we're going to see the pendulum keep shifting about when to initiate therapy for patients. I would certainly consider it and I would talk to him about it, but I wouldn't feel pushed to do it in the clinical setting. I would very much like to see clinical studies of this kind of thing before we start committing our patients, even at this stage, to earlier therapy.

Is there any brand of fish oil that has been found free of mercury?

There are certainly some that are tested. You just need to look at the sources and see what kinds of studies have been done about the fish oil that you might be recommending.

Do we understand what about HIV is causing the low HDL and would it be HIV causing the high triglycerides?

Yes, it's probably HIV that's causing both the low HDL and the triglycerides, and it's probably the inflammation induced by HIV. Any infectious process or inflammatory process can lead to these same kinds of metabolic abnormalities, the low HDL and the high triglycerides. It's just that most of the inflammatory stimuli that people who aren't HIV infected are afflicted with are much briefer, and so they don't have 15 or 18 years of inflammatory stimuli. They have a process that's treated and then abates.

I mentioned atazanavir as an option, what do we know about darunavir and metabolic issues?

Again, I don't think we have the time under our belts to really be able to say whether darunavir is going to offer much of a metabolic advantage.

Is mercury a real problem with fish oil?

I think it's more of a hypothetical problem and risk. Certainly there have been cautions raised about pregnant women eating too much salmon or eating salmon more than once or twice a week. We've actually monitored mercury levels in some of our study patients who get fish oil, and we've never found it to be a problem, but it's always worth thinking about.

Is insulin resistance an adverse effect of using HIV medications?

We know that there may be contribution to insulin resistance by d4T. Indinavir was certainly capable of causing insulin resistance and we don't see that used very much any more. But it's probably more likely that most of the insulin resistance that we're seeing in our patients is related to the accumulation of visceral fat and maybe in our patients who are obese or who are seriously overweight, which we certainly are seeing at this point in time in HIV. I think at this point it's probably less the medications than it is some of the body shape complications and probably for a number of our patients their genetics.

This presentation was also webcast on Dec. 4, 2007 and Dec. 5, 2007.