Published in the Bulletin of Experimental Treatments for AIDS April 1999 issue, by the San Francisco AIDS Foundation.

Selected Highlights from the 6th Conference on Retroviruses and Opportunistic Infections
-- Main Page
-- Clinical Research
-- Protease Inhibitor Side Effects
-- Anti-HIV Drugs and Combinations
-- HIV Transmission and Origins
-- HIV-Related Conditions

April 1999 Table of Contents

Main Page

beta@sfaf.org

Selected Highlights from the 6th Conference on Retroviruses and Opportunistic Infections

Protease Inhibitor Side Effects

By Harvey S. Bartnof, MD

• Benefits of Switching from Protease Inhibitors to NNRTIs
• Fat Redistribution Documented in 69 Persons on Non-Protease Inhibitor Therapy
• HAART-Related Body Fat Redistribution May Be Predicted by Blood Tests
• Anti-HIV Therapy May Decrease Bone Mineral Density

Benefits of Switching from Protease Inhibitors to NNRTIs

• When switching to a nevirapine-containing regimen, some improvement in fat redistribution and blood fat (lipid) levels occurred after three to seven months.
• When switching to a regimen containing efavirenz (Sustiva), abdominal size decreased slightly, weight increased, and blood sugar levels improved; however, blood cholesterol and triglyceride levels increased.

As side effects associated with protease inhibitor therapy continue to increase, some people with undetectable HIV viral loads have expressed interest in switching to regimens based on NNRTIs. The major concerns associated with protease inhibitors center around abnormal increases in blood cholesterol and triglyceride levels and abnormal body fat redistribution, including lipodystrophy. Lipodystrophy refers to the loss of fat in the face, arms, legs, and buttocks, with or without fat accumulation in the abdomen ("protease paunch"), back of the neck ("buffalo hump"), and breasts (see BETA, January 1999). Several presentations at CROI addressed reports of people who made such a regimen switch, usually due to fat redistribution.

One of these presentations was authored by Lidia Ruiz, MD, and colleagues from the Lipodystrophy Study Group in Barcelona, Spain. In order to enroll in this randomized study, HIV positive participants had to have been taking HAART (a protease inhibitor plus d4T/3TC) for at least nine months, have maintained an undetectable HIV viral load (limit of quantitation 400 copies/mL) for at least six months, and have some degree of lipodystrophy. Participants were randomized into one of two arms. Group 1 received d4T/ddI plus the same protease inhibitor. Before entering the study, the subjects' prior protease inhibitor drug therapy included indinavir (74%), nelfinavir (16%), or ritonavir plus saquinavir (10%). Group 2 was given d4T/ddI plus nevirapine. Both groups were evaluated every three months for changes in viral load, lipodystrophy, and blood lipid levels. Body fat changes were measured by dual energy X-ray absorptiometry (DEXA) and bioelectrical impedance analysis (BIA). All who were randomized to nevirapine were given an antihistamine to help prevent rash.

Interim results after three months were reported for 21 of the 60 participants who were enrolled. Ten were randomized into group 1 and 11 were randomized into group 2. HIV viral loads remained undetectable (limit of quantitation 400 copies/mL) for all 21 persons. For the 17 who entered with a baseline viral load of less than 50 copies/mL, all remained undetectable at that level. CD4 cell counts remained greater than 500 cells/mm³ in both arms. The mean blood cholesterol level decreased significantly in group 2 from a mean of 230 mg/dL to 196 mg/dL. Cholesterol levels became normal for seven of 11 persons (64%) in group 2 and one of ten (10%) in group 1.

Blood triglyceride levels decreased in group 2, but not significantly. One person in group 2 had a high fasting blood sugar level both at baseline and at three months. Liver enzyme levels remained normal in both groups. Body shape changes, as determined by both participants and physicians, improved significantly in group 2; on a five-point participant rating scale, body shape changes decreased one point from "high" to "moderate." However, those in group 1 had no change as measured by the five-point scale, and remained "high." The differences were significant. DEXA and BIA measurements confirmed the body shape improvements in group 2, but the difference was not significant. In addition, participants' quality-of-life measurements improved in group 2, often due to a lower daily number of pills.

The authors concluded that three months after switching from a protease inhibitor-based HAART regimen to a nevirapine-based regimen, improvements in blood cholesterol level and body shape can occur while maintaining an undetectable HIV viral load and an increased CD4 cell count. This study is ongoing.

A second report describing a switch from a protease inhibitor-containing regimen to a nevirapine-containing regimen was authored by researchers from the Hospital Clinica, also in Barcelona. A total of 23 persons (52% men) were recruited after becoming aware of body fat redistribution while taking regimens containing one or two protease inhibitors plus two nucleoside analogs. Participants had first noted body shape changes after six to 26 months of protease inhibitor therapy. Increased abdominal size was noted by 78%, while 69% had both increased abdominal size and fat loss in the face, arms, and legs. Only 22% had lipodystrophy only without fat accumulation. HIV viral loads were undetectable (limit of quantitation 200 copies/mL) for a median of nine months before changing from the protease inhibitor-based regimen. The median CD4 cell count before switching was 514 cells/mm³.

After a median of seven months on the nevirapine-based regimen, 91% of the participants reported a partial improvement in their body fat redistribution, particularly in peripheral wasting, although none reported a complete return to pre-protease inhibitor body shape. However, there were no reported clinical measurements of fat redistribution. There were significant improvements in blood cholesterol (decrease of 21%), triglyceride (decrease of 56%), and glucose (decrease of 16%) levels. CD4 cell counts were essentially unchanged. HIV viral loads remained undetectable, with the exception of one person who had a small increase to 546 copies/mL. The authors concluded that switching from a protease inhibitor-based regimen to a nevirapine-based regimen is associated with improvements in both body fat redistribution and blood lipid levels.

The third presentation to address switching from a protease inhibitor triple combination regimen to a nevirapine-containing combination was the "Maintavir" study. F. Raffi, MD, and colleagues from the University Hospital in Nantes, France, enrolled 18 participants (28% women) who wanted to change their protease inhibitor. Five (28%) had a prior AIDS diagnosis. Reasons for switching included a desire for a simpler regimen, lipodystrophy, digestive symptoms, concerns about adherence, and kidney pain. Before the switch, participants were taking indinavir (44%), ritonavir (39%), or nelfinavir (17%). The accompanying nucleoside analogs both before and after the switch included AZT/3TC (67%), d4T/3TC (16%), and ddI/d4T (16%). All 18 participants had been taking the same protease inhibitor regimen for at least one year. Prior to switching, they had maintained an undetectable HIV viral load (limit of quantitation 400 copies/mL) for a mean of 16 months; 61% had a level less than 80 copies/mL.

The protease inhibitor was changed to nevirapine for 89% of participants and to efavirenz for the other 11%. After a mean follow-up of 15 weeks, CD4 cell counts remained unchanged at approximately 530 cells/mm³. HIV RNA viral loads remained undetectable (limit of quantitation 400 copies/mL) for 16 (89%) of participants; one became nonadherent and experienced a viral load increase to 38,000 copies/mL, while another had a detectable viral load of only 95 copies/mL using an ultrasensitive test. Using a test with a cutoff of 80 copies/mL, 55% had an undetectable viral load at follow-up.

The five participants who switched drugs due to lipodystrophy had "subjective and objective improvements in terms of fat accumulation and body modifications." However, no objective measurements were reported. Also, blood fat measurements were not reported. The new regimen was well tolerated by 89% of participants. One person who switched to nevirapine developed a rash but was able to continue therapy. One of three persons who initially switched to efavirenz subsequently switched to nevirapine after three days due to vertigo (spinning sensation).

The fourth presentation addressed switching from a specific protease inhibitor combination to either a nevirapine-containing regimen or a nelfinavir-containing regimen due to body fat redistribution. This approach was tried due to a presentation at the 12th World AIDS Conference last summer indicating that switching from a non-nelfinavir protease inhibitor combination to a nelfinavir-containing regimen was associated with a partial reversal of body fat redistribution in seven of 21 persons (33%) (see BETA, October 1998). Among the remainder of the 21, the condition worsened in 10% and stabilized in the remaining 57% after three months of follow-up.

The report was authored by Andrew Carr, MD, David Cooper, MD, and colleagues from St. Vincent's Hospital in Sydney, Australia. These authors presented their theory of HAART-related lipodystrophy syndrome last summer (see BETA, October 1998). Their CROI poster described 32 persons with HAART-related body fat redistribution who switched therapy to determine whether body shape changes would improve. Sixteen switched from a protease inhibitor to nevirapine while maintaining the other drugs in their regimen. Twelve others switched from either indinavir or ritonavir/saquinavir to nelfinavir. The other four stopped only their protease inhibitor or all of their anti-HIV drugs. Body composition was measured by DEXA.

The results after six months showed that among those who switched to nelfinavir, there were no changes or a slight worsening of both body fat and blood lipid levels. Among those who switched to nevirapine, there were improvements in fasting cholesterol and triglyceride levels, but not levels of HDL ("good") cholesterol. Increased abdominal fat reverted to normal levels. Fat loss in the arms and legs continued to worsen for the first three months, but then improved.

For the 15 participants who switched to nevirapine and had an undetectable HIV viral load (limit of quantitation 400 copies/mL) at the time of the switch, eleven (73%) maintained an undetectable viral load. CD4 cell counts were not reported in the abstract. The authors concluded that a switch to a nevirapine-containing regimen improves fat redistribution and blood lipid levels, but may not always maintain an undetectable viral load. In the authors' experience, switching to nelfinavir did not improve either fat redistribution or blood lipid profiles.

A fifth presentation addressing a similar drug class switch originated in the United Kingdom. Graeme Moyle, MD, and colleagues from Chelsea and Westminster Hospital in London reported on 12 persons experiencing lipodystrophy while on an indinavir-containing regimen for a mean of 21 months who switched to an efavirenz-containing regimen. The participants' nucleoside analogs were not changed, and included d4T/3TC for all but one. Before the switch, all participants but one had an undetectable HIV RNA viral load (limit of quantitation 500 copies/mL). The twelfth had a baseline viral load of 4,834 copies/mL. The mean baseline CD4 cell count was 251 cells/mm³.

In an interim analysis three months after the switch to efavirenz, the mean CD4 cell count had increased to 290 cells/mm³. Viral load remained undetectable in 11 participants and decreased to 857 copies/mL in the twelfth. For the eight persons who reached six months of follow-up, the mean CD4 cell count increased to 342 cells/mm³ and viral loads remained undetectable. The switch from indinavir to efavirenz was reported to be well tolerated. One person changed from efavirenz to nevirapine due to persistent mental concentration problems. Improvements in appearance related to body fat redistribution were reported by most participants. The poster showed baseline and six-month face and leg photographs of one person; the improvements were obvious, and leg hair even reappeared.

The participants' mean waist circumference decreased from 32.2 to 31.8 inches for the twelve who completed three months on the new regimen; five of twelve experienced a decrease of two inches or more. Mean weight increased significantly from 132 to 137 pounds for the twelve persons completing three months on the new regimen, and from 134 to 141 pounds for the eight completing six months. DEXA measurements of total body fat and fat-free mass indicated no significant differences for participants completing either three or six months on the new regimen.

Fasting blood sugar levels improved slightly in six persons. Somewhat surprising was a worsening of fasting blood fat levels after three months that tended to improve towards baseline after six months. Fasting cholesterol levels increased from a mean baseline of 5.9 mmol/L to 7.8 mmol/L after three months, yet decreased to 6.7 mmol/L after six months (normal is 3.5-6.5 mmol/L). Fasting triglyceride levels, while high at baseline, increased to even higher levels on efavirenz. Triglyceride levels increased from a mean baseline of 5.1 mmol/L to 8.0 mmol/L for the twelve persons completing three months on the new regimen, and from 5.9 mmol/L to 7.3 mmol/L for the eight completing six months (normal is 0.5-2.2 mmol/L).

Early interim results suggest that switching from an indinavir-containing combination to an efavirenz-containing combination may be associated with mild weight gain, a slight improvement in "protease paunch," improved blood glucose levels, and a worsening of blood lipid profile. Control of HIV viral load was maintained and CD4 cell counts increased. More participants will be enrolled in this study, and all will be followed for longer periods. Further follow-up will help to elucidate any true changes.

These five presentations appear to have a recurrent theme. Undoubtedly there will be longer follow-up for most of these participants, and new studies will help to define the pros and cons of switching from a protease inhibitor-based regimen to an NNRTI-based combination. As described in the next section, body fat redistribution in HIV positive persons may also be due to non-protease inhibitor anti-HIV drugs.

Fat Redistribution Documented in 69 Persons on Non-Protease Inhibitor Therapy

• Abnormal blood fats also occurred with nucleoside analogs, but to a lesser degree than with protease inhibitors.
• Aspects of lipodystrophy appear more common and worse with d4T compared to AZT.

While protease inhibitors have been most often linked with fat redistribution including lipodystrophy, four presentations at CROI indicated that these side effects can also occur with non-protease inhibitor anti-HIV drugs.

Thierry Saint-Marc, MD, and colleagues from the E. Harriot Hospital in Lyon, France, reported on persons who developed partial or generalized lipodystrophy while taking only two nucleoside analogs. None were taking nor had ever taken a protease inhibitor. The researchers measured lipoatrophy, or loss of fat under the skin in the extremities, in addition to central (abdominal) fat gain. The participants were compared to a control group of 15 treatment-naive HIV positive persons.

The study included 43 HIV positive participants (19% women). The anti-HIV therapy combinations used by the participants were d4T/3TC (33%), d4T/ddI (30%), AZT/ddI (30%), and AZT plus either 3TC or ddC (7%). The researchers divided the participants into two groups: those taking a d4T-containing regimen and those taking an AZT-containing regimen. Virologic and immunologic responses to the double nucleoside analog regimens were good, with a median HIV RNA viral load of approximately three log copies/mL and a median CD4 cell count of 540 cells/mm³. The median duration of exposure to anti-HIV drugs was approximately 1.4 years for the d4T group and 2.1 years for the AZT group. Serial fat redistribution changes were recorded by BIA, computed tomography (CT) scans, and skinfold thickness measurements.

The results showed that total body fat percentage was significantly lower in both nucleoside analog groups (13% fat in the d4T group and 15% fat in the AZT group) than in the control group (17% fat). When comparing the two treatment groups, the differences in total body fat percentages were significant, with the d4T group experiencing more fat loss than the AZT group. The mean amount of fat under the skin in the mid-thigh showed that both treatment groups had significantly less fat (21 cm² in the d4T group and 31 cm² in the AZT group) than the control group (92 cm²). The differences between the two treatment groups were also significant, with the d4T group again faring worse than the AZT group.

The same pattern emerged for the upper arm; the amount of fat under the skin (skinfold thickness) was decreased in both treatment groups (mean 4.9 mm in the d4T group and 5.7 mm in the AZT group) compared to the control group (6.0 mm), with the d4T group losing significantly more fat than the AZT group. The authors also reported that fat loss in the arms and legs was observed during the physician's physical examination in 63% of participants taking d4T, 19% of those taking AZT, and none of the control group. The median time to lipoatrophy in both treatment groups was 14 months, ranging from three to 22 months. In the d4T group, over 90% had evidence of lipoatrophy after approximately 21 months.

Fat loss in the face was not measured. However, the poster included one facial photograph of a person who was taking d4T/3TC. He had obvious fat loss under the skin of the cheeks. No person in the AZT group had visual evidence of facial fat loss.

The results for abdominal fat gain showed the same trend, but in reverse. The mean amount of fat in the abdomen was increased in both treatment groups (103 cm² in the d4T group and 80 cm² in the AZT group) compared to the control group (75 cm²). When comparing the intra-abdominal fat gain between the two treatment groups, the d4T group fared significantly worse (more fat gained) than the AZT group. Blood fat measurements were not reported.

The authors concluded that lipodystrophy in HIV positive persons can occur as a complication of therapy with two nucleoside analog drugs without any protease inhibitor therapy. As has been observed with the protease inhibitor drugs, the occurrence of lipodystrophy associated with d4T is progressive, and its rate increases with increasing duration of therapy. The authors established that several aspects of lipodystrophy are significantly more common and worse with a d4T-containing regimen compared to an AZT-containing regimen. The detailed measurements presented by the researchers adds significant weight to their claims.

The second presentation to address this issue was authored by researchers from the Royal Free Center for HIV Medicine in London. They reported on five HIV positive persons (two men and three women) who developed lipodystrophy syndrome while taking combination antiretroviral therapy without a protease inhibitor. The five were taking various two- or three-drug combinations that included all six FDA-approved nucleoside analogs except ddI. One of the five was taking a triple combination that included nevirapine, an NNRTI. The duration of therapy ranged from ten to 34 months. None had ever taken a protease inhibitor drug.

All five had noted typical body fat redistribution associated with protease inhibitors, but none had a "buffalo hump." All had HIV viral load levels that were very low (3,600 to fewer than 50 copies/mL). The median CD4 cell count was 490 cells/mm³. Interestingly, all five had experienced weight loss since starting anti-HIV therapy (a median loss of 5.2 kg [11.5 pounds], representing 8% of body weight). Blood fat levels including cholesterol and triglycerides were within normal limits.

The authors concluded that lipodystrophy syndrome may be associated with nucleoside analogs and NNRTI drugs without any protease inhibitors. They commented that HIV viral load is unlikely to be related to lipodystrophy due to the very low levels in these five persons.

The third presentation addressing body fat redistribution in the absence of a protease inhibitor was authored by researchers from the Institute of Infectious Disease and Tropical Medicine in Milan, Italy. They reported on 12 women with fat redistribution documented by DEXA. All were taking anti-HIV therapy with two drugs, but none were taking protease inhibitors. In this analysis, body fat redistribution was significantly associated with taking 3TC or d4T. As in the first report above, those taking AZT in this study had a significantly lower risk of fat redistribution.

A fourth presentation measured specific aspects of fat accumulation in women taking anti-HIV therapy. Researchers from Beth Israel Medical Center in New York City asked HIV positive women whether they had experienced an increase in bra size and/or waist size since starting therapy. They compared 95 women taking a protease inhibitor-based combination (group 1) to 32 women taking non-protease inhibitor-based drug regimens (group 2). Approximately one-third in each group reported an increase in bra size (37% in group 1 and 31% in group 2; no statistical difference).

However, for those with an increase, the mean increase was significantly greater in group 1 (1.8 bra size increase) compared to group 2 (0.9 bra size increase). Approximately half of each group reported an increase in waist size (56% in group 1 and 41% in group 2; no statistical difference). For those with a waist size increase, the mean increase was greater in group 1 (3.6 sizes) than in group 2 (2.6 sizes); this difference was not statistically significant. The only two women with a "buffalo hump" were in group 1. The authors concluded that increases in bra/breast size and waist size may occur among HIV positive women who are taking anti-HIV therapy, either with or without a protease inhibitor. However, there were clear trends towards more pronounced changes among those taking a protease inhibitor. Unfortunately, the authors noted that the bra/breast size and waist size increases led to higher rates of poor adherence to therapy.

These four presentations, reporting on 69 HIV positive persons taking non-protease inhibitor-based antiretroviral therapy, strongly suggest that various aspects of body fat redistribution can occur in association with nucleoside analogs. However, there appears to be a trend towards lower rates and lesser magnitude of changes with nucleoside analogs compared to protease inhibitors. Whether different rates are associated with different nucleoside analogs remains to be determined; however, some evidence suggests that d4T may cause higher rates of change. Additional research in this area is ongoing.

HAART-Related Body Fat Redistribution May Be Predicted by Blood Tests

Finding a factor, marker, or profile that could help predict the development of body fat redistribution associated with anti-HIV therapy would benefit people with HIV and their physicians. Australian researchers Carr, Cooper, and colleagues believe that they have found two blood markers that may do just that. A total of 116 HIV positive participants who were taking a protease inhibitor-based regimen for a mean of 21 months were evaluated. Body fat composition measurements were done using DEXA. Blood tests and lipodystrophy measurements were compared to results from eight months earlier.

The researchers determined that fasting levels of triglycerides and C-peptide shortly after starting a protease inhibitor-based regimen were significantly correlated with the severity of lipodystrophy that subsequently developed. The more abnormal the tests were initially, the worse the later lipodystrophy while taking HAART. Also, the two blood tests were able to predict the subsequent absence or presence of mild lipodystrophy. Baseline cholesterol and body fat measurements were not found to be associated with subsequent lipodystrophy. Similar findings by other researchers and in larger numbers of participants are needed before these results can be considered definitive. However, if these findings are upheld, having an early marker to help predict the future development of lipodystrophy would be extremely useful.

Anti-HIV Therapy May Decrease Bone Mineral Density

A small study of 17 HIV positive men from Argentina suggests that decreased bone density may be another side effect of HAART. Such decreases in bone density may predispose people to bone fractures. To date there has not been an increased rate of bone fractures in HIV positive persons, with or without HAART. An increase in hip replacement surgeries among HIV positive persons has been reported at this and other conferences (see BETA, January 1999). However, these were due to avascular necrosis, a type of bone death associated with loss of blood flow at the head of the femur (the ball joint where the thigh bone joins the pelvis).

In the CROI report from the Rawson Hospital in Cordoba, Argentina, the changes in bone density did not occur evenly in all bones. There was a tendency towards sparing of the lumbar (lower) spine and the hip. With the exception of one person, all measurements that may have contributed to decreased bone density were normal, including levels of the minerals calcium and phosphorous and the hormones prolactin and thyroxine. The one person with a predisposing abnormality had a low blood level of cortisol, an adrenal gland hormone.

No person was taking medications known to affect bone metabolism, and none had undergone prolonged bed rest, a factor predisposing to mineral loss. Levels of alkaline phosphatase, an enzyme that may increase with certain bone diseases, were normal in all participants. There was no control group of HIV positive persons taking non-protease inhibitor antiretroviral therapy or not taking any anti-HIV therapy. Additional studies of bone mineral density in HIV positive persons with and without anti-HIV therapy are needed to determine the true rate of this abnormality and its potential causes.

Harvey S. Bartnof, MD, has been a member of the Scientific Advisory Committee of the San Francisco AIDS Foundation since 1987.

Page last updated 1 June 1999