Skip directly to content

2005 Pipeline Report—Part I: Complications

The title not only refers to the multi-side effects that people are dealing with on a regular basis both with and without their meds, but also to the myriad possibilities of what you can take (not counting the drugs out of general use now, nor the two-in-one or the three-in-one formulations, there are 17 drugs to mix and match), what you can't take it with (due to interactions, either within the ARVs themselves or when you are taking non-ARV meds, be they lipid lowerers, antidepressants, OI prophylaxis, or because of access issues), what you try to avoid (due to side effects, previous treatment history), and general treatment fatigue (we are in the ninth year of generalized HAART).

At the 2005 CROI, the 12th Conference on Retroviruses and Opportunistic Infections, we were treated to 774 posters. 92 of these were posters specifically on side effects, ranging from fat redistribution to hepatitis flare-ups. In nearly 100 others, regarding the efficacy of drugs or studies, side effects reporting plays an important (though understated) role. (I am leaving out resistance and its effects.) There were also 182 talks given. Of these, 14 were on adverse events, which maybe underestimates their importance, because all of these 14 talks were given in the largest (2000+ seats) Auditorium in the conference center. At least half of these are webcast, and can be found at along with others like "Can we find better PIs?" "Better" in large part meaning less adverse events. I am not counting other talks, like Controversies, or Challenges, or Responses, also in the Auditorium, which in no small part touched on "complications." Finally, in the pre-conference Workshop for New Investigators, a large chunk of the day (up to 40%) was dedicated to "management issues."

Just say no

Over the past two years, in TAG's—and ATAC's—New Drug Approval Position Papers, many AE issues have been highlighted. A number of these specific issues were revisited in "tolerance" posters at CROI. All posters/abstracts below can be found through putting the poster/abstract number in the search box.

Abacavir (ABC)

FDA presented a poster on the "Increased rate and severity of the hypersensitivity reaction (HSR) to ABC, in randomized controlled clinical trials", especially QD vs BID (abstr 835). An increase was seen simply by crunching numbers—there is more HSR and more severe HSR today than when ABC was first approved. That could be a more stringent vigilance by doctors and patients (chapeau!), but incidence has gone up, be it ABC alone (Ziagen®), in the combo pill with 3TC (Epzicom®), and in the once daily (QD) vs twice daily (BID) administrations. FDA has changed the label and has called for more studies to address this. In poster 836, GSK, with the same data, said HSR is not on the increase, even with the earlier 4.6% measurement increased clearly to 9.6% in the poster (abstr 836). Go figure. They did a wider analysis and saw that African-Americans and males separately had a higher probability of getting HSR, as do people with low CD4.

Tenofovir (TDF)

Does tenofovir accelerating bone loss (abstr 825)? Add TDF to duration of HIV infection and elevated bilirubin (ATV?), and there is "more bone loss." Poster 824 links serum cholesterol concentrations and fasting glycaemia to decreased BMD. Poster 827 talks about Macro CK-2 being an indicator of cellular necrosis and cytotoxicity, and that there are high levels of creatine kinase type 2 (CK-2) in people taking TDF. Poster 823 shows that although there is not much difference between d4T and TDF in bone loss, 75% of people with osteopenia at baseline continue with it at 3 years, while 10% progress to osteoporosis. And most of those with osteoporosis continue having it at 3 years. 10% of those with normal BMD at baseline progress to osteopenia. Posters 818-820 showed an association between tenofovir and low glomular filtration rates (renal dysfunction) and creatinine clearance.

A poster from Poland (abstr 824) suggested that the rates of osteopenia and osteoporosis, regardless of treatment experience, is high in HIV+ people. Osteopenia was observed in 35% of patients, osteoporosis in 19%, while no bone density alterations were observed in the rest. Serum cholesterol concentration (SCC) is associated with decreased bone mass and high osteoblastic activity in HIV+ patients. A further increase of SCC was observed in patients on HAART with even higher values in patients with lipodystrophy. A positive correlation between serum osteocalcin and total cholesterol, LDL cholesterol, triglycerides, and fasting glucose levels were observed in HAART patients, which were not seen in naïve individuals. The decrease in intravertebral marrow fat was negatively correlated with total cholesterol concentrations in HIV+. An increase in serum osteoprotegerin (a marker for osteoporosis) in HIV+ patients vs control was observed. Hypercholesterolaemia may increase the risk of osteopenia and osteoporosis in this population.

Enfuvirtide (T20)

Posters 837 and 838 were by FDA and David Cooper. FDA said that almost all subjects experience a severe grade ISRs which warrants further investigation into improving ISRs (86%). Cooper says that the more peripheral fat you have, the easier time you'll have of it (duh?).

Atazanavir (ATV)

Poster 842, ATV/r decreases lipid accumulation. Plasma lipids also went down in abstr 850. Abstr 858 said it again: lipid profiles improved after 12 weeks on ATV (this time, without /r). Abstr 745 looked at bilirubin as an indicator of adherence. Not perfect.

Other complications

It was shown by various groups that switching from AZT or d4T (thymidine analogs, seen to have negative mtDNA effects) to ABC or TDF switches loss in limb fat to a slight gain. This gain, after 1 year, is approximately 0.5 kilo which is not clinically significant. Of course, the loss has been stopped, and maybe a reversion has started, but it has probably taken a number of years to have lost this amount of fat (through mt depletion and other unspecified mechanisms), and this one year data is a good sign, but the mt effects have not been measured, and the increase in limb fat is not necessarily an "improvement", as Cristine Katlama so elegantly reminded us, it is simply an increase, and very modest at that.

Note: mt (mitochondria) are tiny organelles, scores of which are found within nearly every cell of the human body—they play a key role in releasing cellular energy. This energy flow helps them be both self-replicating, oxidize food compounds and produce ATP (adenosine triphosphate), a phosphorylation of oxygen. Some NRTIs, specifically thymidine analogs, cause mitochondria's amino acid(s) to act slightly differently. The affected protein(s) still work, but inefficiently, thus, possibly, many organ-specific pathological changes may be related to the defective mtDNA replication (myopathy, neuropathy, lactic acidosis, congestive heart failure, hepatotoxicity, metabolic abnormalities).

Christine Katlama (France) and Pablo Tebas (Spain) got into a sparring match over whether NRTI-sparing regimens offered an increase or an improvement in peripheral fat. Tebas (abstr 40, reporting for ACTG5125) was gentleman enough to give in, that in fact, the reversal of fat loss in the arms and legs was, at least at one year, an increase that was only the beginning of an improvement. Thanks for coming from Paris to correct the ACTG's English, Christine. Although, the same no-NRTI regimen (Kal+EFV) thno-PI arm.

Although there were many studies that showed that switching from AZT or d4T-based regimens to ABC or TDF based ones improved fat loss (Graeme Moyle with the RAVE study, Rob Murphy with ACTG 5110, Pablo Tebas, MITOX), what was gained probably can't yet be called an improvement. At one year, the peripheral fat loss was stopped and a non-clinically significant increase was measurable. So, the new drugs switched to (whether they are NRTIs, or even an NRTI-sparing regimen (Kal+EFV, Kal+NVP)) helped stop one problem but the significance is not clear (is the mtDNA depletion halted? Is the virological failure larger with no NRTIs?). Moyle figured that the reversal after one year is approximately 12% of what was lost (so, another 6 years on this therapy and your peripheral fat will be returned?) Actually, in these NRTI sparing regimens, although limb fat loss was halted and begun to be turned around, lipid issues (as would be expected with Kaletra) worsened.

In two of the studies looking at sparing the NRTIs to help the lipoatrophy, Moyle (abstr 44LB) and Murphy (abstr 45LB) showed that sparing the NRTI class or at least avoiding thymidine analogs (AZT, d4T) may be helpful strategies. Both of these studies had approximately 100 people with more than 500 CD4s. Murphy's had 600 CD4s. A break from treatment period (even a short one) might have done just as well. Maybe in the next trial, there can be an arm of just that—an arm that continues, an arm that spares NRTIs, an arm that spares AZT/d4T, and an arm that stops (in people who have more than 500 CD4s, for example). Moyle did note that ABC had more discontinuations and side effects than TDF (more than 6% hypersensitivity).

Peter Reiss from Amsterdam talked of Lipodystrophy as if it were a Puzzle, doing an overview (abstr 65 and webcast at that succinctly gets one up-to-date regarding the most important issues—including the mitochondrial effects of many of the NRTIs and laying out the many things that are still not clear about changes in body fat distribution. He mentions that it is not easily reversible, and goes as far as to suggest that even without knowing the precise mechanisms, we have a decent idea of those drugs more associated with it and so maybe more avoidable for a reduced (or at least delayed) incidence.

Kate Mulligan showed mixed patterns of change from ACTG substudy 5005S (abstr 38), which looked at 2 nukes and 1 NNRTI. While about one-third gained fat in both trunk and limbs, another third lost fat in both areas. What we consider the "typical" gain of fat in the trunk and loss in the limbs was seen only in about a quarter of patients. She published just a few weeks earlier (Jan 2005) in AIDS a look at a part of the WIHS cohort, in this substudy, some 271 women. In three groups, HIV-, HIV+ on HAART and HIV+ not on HAART, she saw a significant loss of leg fat in the women on HAART, with a background of generalized obesity in all three arms. There also was a tendency in the HIV+ PI group to retain trunk fat vs the HIV+ no-PI group. She mentions that these women are comparable to the initial report from Australia on lean white men. Also, nadir CD4 count correlates with leg fat in these women. Age >30 is associated with more fat in this study. Exercise warrants further investigation.

A small French study (abstr 39) showed a positive effect on lowering triglycerides of an omega-3 capsule taken TID for 8 weeks.

Rosiglitazone was once considered to be a possible help for lipid problems. It has been shown not to be of help, and one more time by a Toronto study of lipid profiles and insulin resistance (abstr 854) in 96 patients over 24 weeks. Looking at why this might be, various groups from Australia saw that it might have to do with previous thymidine analog use (abstr 41) that shuts down expression of just what the rosiglitazone is active on, PPAR gamma.

How good is good?

A number of studies looked at adverse events in children (abstr 53—56, 99, 410, 978, 760—778) and in women (NVP in women, abstr 784), and many studies looked at them as factors in predispositions, like abstr 42, Effect of sex, age and lipid changes, a DAD analysis presented by Dr Wafaa El-Sadr. Looking to see if the MI effect of HAART continues, she presented an analysis of their 23,441 patients (24% women).

By 2004, the myocardial infarction incidence increased from 1.39/1000 patient-years in those not exposed to HAART, to 2.53/1000 patient-years in those exposed for < 1 year, to 6.07/1000 patient-years in those exposed for = 6 years. After adjustment for other potential risk factors, there was a 1.17-fold increased risk of myocardial infarction per additional year of HAART exposure. While the overall absolute risk of myocardial infarction remains modest, the risk continues to increase with longer exposure to HAART over the first 7 years of use. Although the rate of myocardial infarction was higher in men than women (2.04), the relative risk associated with HAART was similar in men (1.14) and women (1.38). The relationship was similar in younger and older patients (men > 45 and women > 55 years). Including time-updated levels of serum total cholesterol, HDL cholesterol, and triglycerides in the same model reduced the association of an additional year of HAART with myocardial infarction to 1.10. Lipodystrophy was not associated with the risk of myocardial infarction. Dyslipidemia explained part but not all of the association of HAART with risk of myocardial infarction.

Docs from my barrio

I had the luck to interview two clinicians from Barcelona, both of whom focus greatly on adverse events. Ana Milinkovic and Esteban Martinez work as a team in the Hospital Clinico in Barcelona.

Martinez spoke on the relative CV and lipid risks that different HIV drugs might present (abstr 63). Although lipid "derangements" may be related to HIV infection itself, they are far more common and severe in association with HAART, irrespective of the drugs used. PIs/r induce more lipid abnormalities than non-boosted PIs and stavudine has more severe effects than other nucleoside reverse transcriptase inhibitors (NRTIs). Ritonavir has the strongest hyperlipidemic effect among PIs while atazanavir appears to have little if any. One compounding issue is that even with similar HAART, lipid disorders are more common in antiretroviral-experienced patients, with prior history of hyperlipidemia, or with clinically evident body fat changes. Diet has not been consistently related to lipid abnormalities in HIV+ patients, although the nutrition industry may take issue. Physical exercise may improve hypertriglyceridemia, which again in the literature, may be more compelling than the 16-week data that shown to be of limited effect in HIV. Switching from PIs to NNRTIs or abacavir or atazanavir, as well as from stavudine to tenofovir may improve lipid abnormalities, although distinct adverse effects and virological failure may arise. Dosage reductions in boosted PIs or stavudine and therapy interruptions may affect lipid parameters positively but their impact on the control of HIV infection is not clear. Fibrates, low-dose (10 mg) atorvastatin, or both may be used but their lipid-lowering effects are usually lower than in the general population.

It seems reasonable to evaluate lipid disorders in HIV+ patients according to the same criteria used in the general population, in the absence of HIV-specific data. The impact of individual antiretroviral drugs on lipid parameters should be included among the factors to be considered on prescribing HAART.

Although Martinez's major poster was on Spain, the results may be applicable across cultures (abstr 870). To wit, in cardiovascular disease, which the DAD cohort has now given a risk factor of 6.70/1000 pt years to (see above), there are some common factors that may lower the risk, some of which are adjustable, like smoking, obesity (including diabetes and blood pressure as well as waist circumference), cholesterol and triglycerides, which can be worked on to some degree, and some of which are not adjustable: sex, age, family history of disease. Tobacco smoking, and high plasma levels of total cholesterol and triglycerides were identified as the factors more heavily associated with a high cardiovascular risk in HIV+ patients in Spain, all of which can be dealt with, although some things, like changing diet and going to the gym may be more long term than immediate (so start already!). The DAD cohort has not seen that lipid-lowering agents have any demonstrable effect. In a related paper in the NEJM in December 2004, Hu showed that higher weight and less physical activity were both shown to be independent predictors of death (in 100,000 American nurses over 24 years). If you feel well enough, get moving! From the accompanying editorial: be fit and lean if you can be!

Milinkovic spoke of the normalization of killer inhibitory receptors and lectin like receptors, but not natural killer cells, in patients on HAART (abstr 513). Natural killers drop after one year on an immunologically-driven STI, but not in a virally-driven one. CD4 cells are lower after one year of an ID STI than a VD STI or the (on treatment) control. After one year, there was no difference in any of the 3 arms.

Milinkovic also presented a poster on reducing the dose of d4T to 30 mg as the normal maximum dose (abstr 857), and that 40mg is effectively an overdose which may be part of the problem all these years. She compared that to d4T(40) (standard dose) vs TDF. Median CD4s were 612 (again, my quick answer, is get these people off treatment if they are suffering mt issues, etc). One person in the d4T(40) arm got hyperlactataemia. Trigs continued to rise in the d4T(40) group (+19), they fell by 40 in the d4T (30) arm, and fell by 133 in the TDF arm. Total chol went up by 4 with d4T(40), down by 4 with d4T(30) or down by 28 with TDF. Total fat and limb fat followed these indicators, with d4T40 (-597, -247), d4T30 (+332, +77), and TDF (+1005, +440).

Hyperlactatemia (increased plasma lactate) with a normal blood pH is lactic acidemia. Lactic acidosis consists of a high lactate level accompanied by metabolic acidosis. The spectrum of lactic acidemia ranges from fulminant multiorgan dysfunction—severe acidosis and hemodynamic instability, to less severe symptomatic hyperlactatemia with hepatic steatosis (fatty liver), to intermittent or chronic low-grade hyperlactatemia without acidosis, steatosis, or symptoms.

Although most cases of lactic acidemia are asymptomatic, a variety of nonspecific presenting complaints have been described, including nausea, vomiting, and diffuse abdominal pain; fatigue, weakness, weight loss, rapid or difficult respiration on exertion, heartbeat disturbances, and neurologic findings have also been reported in the absence of gastrointestinal complaints. Liver abnormalities, including hepatomegaly (enlargement of the liver), hepatic steatosis, and elevated serum transaminases are common in symptomatic hyperlactatemia and almost ubiquitous in NRTI-induced lactic acidosis. The onset of symptoms is usually not acute, occurring over weeks to months, although acute fulminant cases associated with multiorgan (especially liver) dysfunction occur rarely.

Where do we go from here? As the complications get better and better defined, we want to steer clear of certain drugs, which can't always happen, and certainly not in any speedy fashion. It took a good while to bury ddC (the sponsor just announced its discontinuation in March 2005!), while some 50% of people not in the first world are still taking d4T at full (over-) dose, and need to switch out (but to what, and how?). Treatment Interruptions done in conjunction with your doctor are becoming more appetizing, at least in the short term, for short periods of time, which may help in the complications arena. More generalized healthy attitudes need to be taken up by the ever-greying HIV+ population that is feeling these concerns most. More and better lifestyle studies are needed, specific to the different HIV+ populations.

Metabolic Complications of HIV Therapy, HIV InSite Knowledge Base Chapter, May 2003, D C.Chow, et al, HIVInsite.

Am J Physiol Heart Circ Physiol 284: H1-H9, 2003; Vol. 284, Issue 1, H1-H9, January 2003, Defective mitochondrial DNA replication and NRTIs: pathophysiological implications in AIDS cardiomyopathy, W Lewis, Department of Pathology, Emory University, Atlanta, Georgia.

Although not much was presented on azones and fibrates during CROI, a helpful paper came out right after: Pioglitazone improves HAART-related "metabolic syndrome", while fenofibrate was not shown to be useful. Studies in the past of rosiglitazone in patients on HAART find the drug improves insulin resistance, but increase triglyceride levels.

Among patients on pioglitazone for 12 months, mean insulin resistance index values, quantitative insulin sensitivity check index and fasting insulin levels improved significantly. Pioglitazone increased mean HDL cholesterol levels, and decreased mean triglyceride levels by 27.5%. While the drug tended to increase LDL cholesterol as well, LDL particle size also tended to increase, "which resulted in lower levels of atherogenic LDL cholesterol."

No significant effect of fenofibrate was seen on insulin resistance, lipid levels or blood pressure.

A Harding, Reuters Health, 2 March

A Gavrila, Beth Israel Deaconess Medical Center, et al, 1 Mar 2005, Clin Infect Dis 2005;40:745-749.

Kind thanks to AIDS Action Baltimore for permission to reprint this article.