Published in the Bulletin of Experimental Treatments for AIDS October 1998 issue, by the San Francisco AIDS Foundation.

Part 1:
-- Top 40 Headlines from the World AIDS Conference
-- Trial Analysis
-- Viral Load And DNA Measurement
-- Immune-Based Therapy
-- Protease Inhibitor Side Effects
Part 2:
-- Anti-HIV Drugs and Combinations
Part 3:
-- Salvage Therapies
-- Protease Inhibitor Combinations
-- Opportunistic Infections And HIV-Related Conditions
-- HIV Transmission
-- HIV Prevention

October 1998 Table of Contents

Main Page

beta@sfaf.org

Selected Highlights from the 12th World AIDS Conference -- Part 1

Harvey S. Bartnof, M.D.

Over 13,600 participants from 136 countries attended the 12th World AIDS Conference in Geneva, Switzerland, held June 28-July 3, 1998. The theme this year was "Bridging the Gap," referring to the wide gap between developed and developing countries in access to HIV therapies.

Over 6,250 abstracts were presented, a 20% increase over the last international AIDS conference in Vancouver two years ago. There were four main tracks: Basic Science; Clinical Science and Care; Epidemiology, Prevention and Public Health; and Social and Behavioral Science.

Only 200 physician researchers and five community representatives were allowed to participate in the 2nd International Workshop on HIV Drug Resistance and Treatment Strategies, which took place at Lake Maggiore, Italy, on June 24-27, 1998. The workshop is regarded as a prominent forum in the forefront of research on HIV/AIDS treatments and drug resistance. One hundred fifty-three abstracts were presented.

1. Each day, 16,000 people worldwide become infected with HIV; 40% are women, 10% are infants, and 95% live in developing countries.

2. Protease-sparing efavirenz (Sustiva) triple combination therapy is more potent than standard indinavir-containing triple combination at 24 weeks.

3. Abacavir (Ziagen) triple combination with two other nucleoside analogs may be as effective as standard protease inhibitor-containing triple-combination therapies.

4. Adefovir (Preveon) triple or quadruple combinations may be as effective as standard indinavir-containing triple-combination therapy.

5. Delavirdine (Rescriptor) triple combination with two nucleoside analogs may be as effective as standard protease inhibitor triple combinations at 52 weeks.

6. Salvage regimen with four drugs (nelfinavir/saquinavir/nevirapine/abacavir)

is effective up to 20 weeks for those for whom indinavir has failed.

7. The new goal of anti-HIV therapy is an HIV RNA viral load of less than 50 copies/mL.

8. HAART plus the Remune vaccine leads to better HIV immune responses than HAART alone.

9. Researchers call for standardized reporting of drug trial results.

10. Response to HAART correlates with presence and quantity of thymus gland tissue; thymus transplant may play a role in immune reconstitution.

11. International AIDS Society (IAS-USA) and U.S. federal government both publish updated HIV treatment guidelines.

12. IAS-USA guidelines for HIV drug resistance testing are published in the Journal of the American Medical Association.

13. Measuring blood levels of anti-HIV drugs may have clinical benefits.

14. Lipodystrophy rate of 84% occurs among patients taking protease inhibitor therapy (12% of cases are severe); lipodystrophy hypothesis is published in The Lancet and a definition of the syndrome is proposed.

15. Premature heart attack, blood clots in the veins, and gout are newly reported side effects from protease inhibitor drugs; researchers recommend that treatment for high blood fat levels should be considered.

16. Researchers propose that all patients taking a protease inhibitor should receive an oral glucose tolerance test, since 23% have abnormal glucose metabolism (including 15% with diabetes).

17. "Buffalo hump" and "protease paunch" are partially reversed by human growth hormone (Serostim) in three patients.

18. Adefovir is associated with a 32% rate of kidney toxicity after 48 weeks in treatment-experienced patients; persons taking adefovir should receive monthly tests of blood creatinine and salts.

19. Peripheral neuropathy is improved with recombinant human nerve growth factor.

20. Elective cesarean section plus standard AZT regimen decreases mother-to-child HIV transmission to less than 3%.

21. Birth defects may be due to triple-combination anti-HIV therapy during pregnancy.

22. Women have lower HIV viral loads than men, even after adjusting for age and CD4 counts; this has major implications for treatment guidelines.

23. Nineteen percent of HIV positive women in a U.S. study were deficient in vitamin A; the deficiency increases risk of HIV transmission and progression.

24. Vitamin A deficiency in HIV positive women is a better predictor of cervical dysplasia than CD4 count.

25. HAART decreases the rate of cervical squamous intraepithelial lesions in women with HIV.

26. Widespread use of anal Pap smears is calculated to be cost effective and would save lives.

27. AIDS-related brain lymphoma responds to ganciclovir plus interleukin 2 and HAART; HAART improves lymphoma in three patients.

28. HAART decreases the rate of tuberculosis by 50% in São Paulo, Brazil in 1997.

29. John G. Bartlett, MD, of Johns Hopkins University declares HAART to be one of the most cost-effective therapies of the past decade.

30. A gay man from San Francisco and eight others from the U.S. and Switzerland were infected with an HIV strain that was resistant to all four approved protease inhibitor drugs.

31. The HIV strain most common in developing countries has a much higher heterosexual transmission rate than the strain currently most common in the U.S.

32. Resistance to HIV infection is better understood in sex workers and their relatives in developing countries.

33. Lower rate of sexually transmitted diseases was measured in Thai female sex workers with access to both female (internal) and male (penile) condoms.

34. Seven percent of surveyed HIV-negative gay/bisexual men in six U.S. cities used "female" condoms for receptive anal sex during a recent 6-month period.

35. FDA approves latex dental dams for oral/vaginal sex.

36. Phase III trials of AIDSVAX preventive vaccine begin; trials will include 5,000 volunteers in five U.S. cities (Chicago, Denver, Los Angeles, Philadelphia, and St. Louis) and 2,500 volunteers in Thailand.

37. Four pharmaceutical companies (GlaxoWellcome, Hoffmann-La Roche, Bristol-Myers Squibb, and Organon Teknika) decrease the price of HIV therapies for Uganda, Côte d’Ivoire, Vietnam, and Chile.

38. Specific HIV prevention programs decrease sexually transmitted disease rates among disadvantaged racial/ethnic populations in the U.S.

39. Acceptance of homosexual identity is associated with decreased HIV risk behavior among men who have sex with men.

40. Medicaid coverage of anti-HIV therapies for all people with HIV would be cost effective and would save lives.

Two researchers, in trying to decrease much confusion in the arena of HIV/AIDS research, have called for standardized reporting of the results of drug studies. Not having a standard makes interpretation of trial results and comparison of various drugs much more difficult for researchers and physicians (including this reviewer).

It would appear that much, but not all, of the blame for the problem may fall on the statisticians at the various pharmaceutical companies. Also partly to blame may be the media relations personnel at these companies, who have an obvious incentive to optimize the "spin" so that a particular drug will appear to have an advantage over another drug.

Melanie Thompson, MD, from the AIDS Research Consortium of Atlanta and a member of the International AIDS Society treatment guidelines panel, addressed the issue at a July 1998 satellite symposium of the Geneva conference sponsored by Roche. Andrew Hill, MD, from Glaxo Wellcome and representing the AVANTI trial researchers, authored abstracts presented at the AIDS conference and at the Lake Maggiore workshop. Both researchers outlined examples of misleading reporting of drug trial results due to communication that is not standardized.

The best example is reporting of viral load undetectability, a common endpoint in the area of HIV/AIDS therapies. The first misleading factor is the variable cut-off for the limit of detection for viral load measurements. There are numerous studies in which the level of undetectability is 100% using the first generation viral load test (which has a lower limit of 5,000-10,000 copies/mL), but is only 85% using a second generation test (with a lower limit of 400-500 copies/mL). Using the third generation test with a lower limit of 20-80 copies/mL, the same study might show a level of undetectability of 50-65%. Therefore, if study A reports undetectability using a second generation test and study B reports undetectability using a third generation test, the results of the two drug trial results cannot be directly compared.

A second misleading factor in communicating drug trial results is in the number of patients originally assigned that are included in the analysis. If people "drop out," "do not follow up," or experience severe adverse events and discontinue the drug, sometimes researchers do not include them in the final analysis of the results. If only those who are still in the trial taking drugs are used to calculate results, this can obviously inflate the benefits of the treatment.

The strictest method of analysis is called "intent-to-treat." This means that all patients initially assigned to a particular arm of a study, or who were "intended to be treated," are included in the analysis. A less strict analysis is the "as-treated" or "on-treatment" method, which means that only those patients who are still being treated are included in the analysis; drop-outs are excluded. Other statistical "tricks" include so-called "censoring" methods, whereby certain types of patients who drop out of the study may be excluded from the final analysis.

Another statistical method that can lead to confusion falls in the area of interim analyses. At many of the major HIV/AIDS conferences, trials are not yet complete. However, university researchers need to publish and drug companies want to sell their products. Therefore, communicating interim results is quite common. Often only those patients who have reached a particular time point are included in the analysis. In other cases, earlier results from patients who have not reached the interim time point are also included in the analysis (so-called "last observation carried forward"). The latter method may inflate (or sometimes deflate) the benefits of a treatment because an earlier viral load result may be higher (or sometimes lower) than that of those who reached the designated time point.

Another statistical trick is to communicate "peak" or maximal viral load reductions, rather than the viral load reduction at a certain time point. Depending upon which generation viral load test is used, the same study may have a peak viral load reduction that can vary by up to 100-fold.

A last area that makes comparing different trial results difficult is specific characteristics of the trial participants. As a general rule, HIV positive people who have never taken any previous anti-HIV therapies do much better than those who have taken previous drug treatments. Those who have had significant exposure to previous drug therapies often do not have a durable viral load response to many new combinations. Sociodemographic characteristics may also affect trial results. Therefore, meaningful comparisons of drug trial results can only be made when the participants are essentially equivalent.

Hopefully, the research and pharmaceutical industries will heed the recommendations of Thompson and Hill, and will seek to communicate drug trial results in a standardized fashion. The IAS-USA and the U.S. Public Health Service could publish a set of recommendations for standardized reporting of HIV/AIDS drug trial results. This would represent a win-win situation for all -- especially the patients.

One of the more significant changes in the daily care of people with HIV/AIDS as a result of the Geneva conference is the recommendation that the goal of anti-HIV therapy should be to achieve an HIV RNA viral load of less than 50 copies/mL. This represents a substantial decrease from earlier guidelines. The recommendation was made by the International AIDS Society (IAS-USA) at their symposium and published in the July 1, 1998 special HIV/AIDS issue of the Journal of the American Medical Association. The IAS-USA guidelines state, "The most sensitive [viral load] assays available are thus recommended for continued monitoring of response to therapy." However, the guidelines also note that "a baseline plasma HIV RNA level obtained using the more sensitive assays is not generally needed, as more routinely available standard assays will suffice." In addition to the IAS-USA panel of prominent HIV/AIDS researchers, others who agree with the 50 copies/mL viral load goal include John Mellors, MD, from the University of Pittsburgh, and David Cooper, MD, from the University of New South Wales in Sydney, Australia.

The reason for the change is that the lower the viral load can be pushed using anti-HIV therapy, the more durable the viral load response is. Stated another way, people who achieve a viral load of less than 50 copies/mL remain at that level for significantly longer than patients who achieve a viral load of between 50 and 500 copies/mL or higher. This occurs because those who achieve a viral load of 50-500 copies/mL experience ongoing HIV replication, and their remaining virus is likely to develop resistance to the drugs they are using.

One of the reasons for the new recommendation is the imminent availability of new "third generation" viral load tests which can measure as low as 20-80 copies/mL (see table). As of this writing, none of the new tests are yet approved by the Food and Drug Administration (FDA).

The new recommendation means that other guidelines for viral load testing must also be modified. These include:

1. "Viral load tests should be done every four weeks for the first four months of therapy to determine whether the decline has plateaued or rebounded 1 log or more from the nadir [lowest point]," according to Mellors.

2. Two weeks after starting or changing therapy, expect a 1-log (10-fold) decline in viral load every 7-10 days.

3. It may take longer than 24 weeks for an undetectable viral load to be attained using a test with a 50 copies/mL limit of detection, compared to a test with a 500 copies/mL cut-off.

4. If other treatment options exist, change or intensify regimens that do not reduce plasma viral load to less than 50 copies/mL by 16-24 weeks.

5. Frequency of viral load monitoring may need to be increased (e.g., every two months) when using more sensitive tests in order to detect early viral rebound when re-establishment of control of viral replication is more likely.

6. Before switching therapy due to viral load increase, confirm the result with a repeat test.

7. When screening new patients and following patients who are untreated, or for whom treatment is failing, use the second generation assays with a limit of detection of 500-500,000 copies/mL.

Mellors cautioned that, "The minimal change in HIV viral load that is biologically significant is uncertain, particularly at low copy levels...assay variation may result in intermittently detectable virus."

While the lower viral load goal is understandable with respect to the potential long-term implications of failing to suppress HIV replication 100%, it is possible that it may create frustration among patients and physicians. Approximately 50% of patients do not achieve HIV suppression to below the 500 copies/mL limit of detection of the earlier test. Failure to suppress HIV to below 50 copies/mL will certainly occur in many more than half of people treated. Yet to date, there has been no evidence of associated clinical progression in the short term (up to approximately two years) in people who continue treatment with a detectable yet very low viral load. It remains to be seen whether a person who maintains a viral load of, for example, 250 copies/mL for ten or more years will fare any worse than a person who maintains for ten years an undetectable viral load using a test with the 50 copies/mL cut-off.

New Intracellular HIV RNA Research Test

• A new intracellular HIV RNA viral load test reveals abnormal increases inside T-cells before they can be seen on standard blood plasma tests.
• The ultrasensitive fluorescence in situ hybridization test has a limit of detection of 20-30 copies of HIV RNA per cell tested.
• The test measures unspliced messenger RNA transcripts from the HIV gag-pol genes.
• The new test could eventually be used to monitor patients whose HIV viral load becomes undetectable using the standard blood tests.

After 2.5 Years of HAART, HIV Can Still be Grown from Lymph Nodes

• Analysis was done on lymph node tissue from a single man whose treatment included indinavir plus two reverse transcriptase inhibitors.
• The man's HIV RNA viral load had decreased to less than 20 copies/mL.
• HIV proviral DNA was detected in 7 of 100,000 lymph node CD4 cells after two years of HAART.
• The HIV strain found was resistant to reverse transcriptase inhibitors but sensitive to protease inhibitors.
• HIV proviral DNA was not detected in CD4 cells in the blood.

With the exception of HIV long-term non-progressors, most people with HIV have inadequate lymphocyte proliferation responses to various HIV antigens, even after two years of successful response to HAART. In a late-breaker presentation, Fred Valentine, MD, from New York University, presented preliminary results from a study indicating that HAART plus Remune immunogen (a therapeutic immune booster vaccine) improves such responses.

Remune is composed of inactivated whole HIV-1 virus particles with all gp-120 (an envelope protein) removed. It is a type of therapeutic (not preventive) vaccine initially conceptualized by the late Jonas Salk, MD, developer of the currently used polio vaccine.

A total of 43 HIV positive participants were randomized into two groups. The mean baseline viral load was 8,159 copies/mL and the mean baseline CD4 count was 493 cells/mm³. Both arms received indinavir plus AZT/3TC; the abstract did not indicate whether they had prior anti-HIV therapy. One arm also received Remune injections at weeks 4, 16, and 28, while the other arm received placebo injections. Both groups will be followed and evaluated at week 32 for lymphocyte proliferation responses to various HIV antigens.

By week 16, 86% (18 of 21) of the HAART/Remune arm achieved an undetectable viral load (limit 40 copies/mL), compared to only 67% (12 of 18) in the HAART only arm.

At week 20, the mean CD4 cell increase was 137 cells/mm³ and 109 cells/mm³, respectively, for the two arms. The HAART/Remune arm had significantly increased lymphocyte proliferation responses to three different HIV antigens in vitro compared to baseline. The HAART-only arm did not develop any responses to the same HIV antigens. Some participants in the Remune arm developed lymphocyte response levels similar to those of long-term non-progressors. This interesting study will continue.

In another report, monthly injections of Remune for four months were given to 30 HIV positive volunteers in Thailand in an open-label study. The E strain of HIV is predominant in Thailand, while the B strain is most common in the U.S. and Europe. Additional injections were given at weeks 24, 36, 48, and 60. No other anti-HIV therapies were used. Follow-up continued for two years. When compared to baseline levels, CD4 cell percentage and body weight increased significantly, and HIV RNA viral load decreased significantly. In addition, Western Blot antibody test results showed an increase in the number (repertoire) of HIV antibodies in 28 persons, suggesting a stronger immune response. No progression of HIV/AIDS was observed. The vaccine was "tolerated well with no serious adverse events reported."

Another report described an open-label U.S. trial with 11 HIV positive volunteers. Remune injections were given every three months and patients were followed for just over one year. In addition to a significant increase in lymphocyte responses to HIV antigens, participants also showed a significant decrease in tumor necrosis factor-alpha, a cytokine (chemical messenger) that is abnormally elevated in people with HIV disease. Levels of the chemokine RANTES were also significantly increased; increased RANTES levels are believed to be associated with a slower rate of HIV disease progression.

Thymus Transplant Shows Good Results

• Thymus gland transplant plus HAART leads to "successful immune reconstitution" in 13-year-old boy.
• Response to HAART is linked with amount of thymus tissue present.
• Research on immunodeficient mice suggests benefit from transplant of thymus tissue from other species.

A few abstracts presented in Geneva revealed new insights into the role the thymus gland may play in HIV/AIDS disease and immune restoration. In HIV negative persons, the thymus gland, located behind the breastbone, shrinks during the teenage years. In adult life, it is thought to play a minor role (if any) in normal immune function.

Researchers from Emory University have reported "successful immune reconstitution" in a 13-year-old boy with AIDS who received a human thymus transplant and HAART. The boy was infected with HIV when he was one year old. Before the surgery, he had been taking AZT/3TC. He experienced weight loss, several opportunistic infections and a very low CD4 count, even though he his viral load was undetectable. Over two years after the transplant and beginning HAART, he was without illness, had a CD4 count of approximately 250 cells/mm³ , and still had an undetectable viral load. In laboratory tests, he showed T-cell responses to several foreign antigens and a normal T-cell repertoire of T-cells (restoration of the T-cell repertoire in adults with AIDS who respond to HAART is slow and partial). The boy had no complications from graft versus host disease (a condition in which transplanted immune tissue attacks the new host). The authors believe that the transplant was a source of T-cell clones needed to respond to new foreign antigens.

In another report, researchers from Rush Medical College in Chicago attempted to determine the relationship between CD4 cell increases after HAART and baseline quantity of thymus gland tissue. A total of 30 persons were evaluated before and after 48 weeks of HAART (ritonavir/AZT/3TC). Chest computerized tomography scans were performed at baseline and at week 48, and radiologists rated the amount of thymus tissue as either "minimal" or "moderate-abundant." Persons with a greater mass of thymus tissue had significantly higher CD4 cell counts at baseline and after 48 weeks of HAART. The cells were of the "naive" type required to mount an immune response to new foreign antigens. However, increases in naive CD4 cell numbers was not significantly associated with thymic mass. Even those patients with minimal thymus mass at baseline had significant increases in naive CD4 cells.

A similar report from Rome, Italy, found that baseline thymus volume in children treated with indinavir/d4T/3TC for 24 months was positively correlated with immune function improvement including CD4 cell count increases, repertoire, and näive T-cell increases. The authors concluded that the thymus gland may play an important role in immune reconstitution in HIV disease. Finally, researchers from Harvard University and the National Institute of Allergy and Infectious Diseases used thymus transplantation in mice with genetic immune deficiency to better understand the immunology of HIV disease progression. Severely immunodeficient mice with transplanted pig thymus and human liver tissue survived and showed evidence of CD4 cell reconstitution after being injected with HIV; in contrast, an identical HIV injection into mice with human thymus and liver tissue transplants developed AIDS-like illness and severe T cell depletion.

The results indicate that a thymus gland transplant from another species (e.g., pigs, which do not become ill after HIV injection) may have a role in immune reconstitution in HIV/AIDS.

Heart Attack, Angina and High Blood Fat Levels

• Two young men experience heart pain; one suffers heart attack.
• Chart review shows that high blood fat levels are common in people taking protease inhibitors.

Conference attendees discussed the recent report of coronary artery disease ("hardening" and blockage of blood vessels supplying the heart muscle) in two young men with AIDS from Minnesota. Both were taking a protease inhibitor and had marked increases in their blood fat (cholesterol and triglyceride) levels. Clinicians have been concerned about the possibility of heart disease and stroke due to the extremely high blood fat levels that can result from protease inhibitor therapy.

The first man was a 26-year-old smoker and was prescribed ritonavir/saquinavir/d4T/3TC because of a high viral load of more than one million copies/mL and a CD4 count of 10 cells/mm³. After only four weeks of HAART, he was admitted to the hospital and found to have had an acute heart attack (myocardial infarction), confirmed by angiography (dye studies).

The second man was a 37-year-old diabetic; he developed heart pain (angina) while shoveling snow. His HAART regimen included indinavir. Angiography revealed severe coronary artery disease with partial blockage in two of the three main arteries.

In a chart review, Keith Henry, MD, and colleagues found that 33% of 124 clinic patients taking a protease inhibitor had a blood fat abnormality that would trigger intervention according to the current guidelines of the National Cholesterol Education Program. Many of the 41 patients implemented different interventions to manage their high fat levels, including a diet and exercise regimen. Fifteen patients started therapy with gemfibrozil (Lopid), a fat-lowering drug. After two months, the mean cholesterol level decreased by 25%, while the mean triglyceride level decreased by 33%.

This reports and others appear to shift the risk-benefit ratio for protease inhibitor therapy in patients with HIV/AIDS. The risks are greater than initially appreciated. What good is an undetectable HIV viral load if someone dies from an early heart attack due to side effects of protease inhibitor therapy?

Researchers from Iowa City, Iowa and Omaha, Nebraska, measured the rate of severe blood clots in their HIV positive patients. Before the advent of protease inhibitors, only one patient out of 660 in eight years developed such a clot. However, since protease inhibitors have been available, six out of 783 patients (16% of them women) have developed a total of 11 life-threatening blood clots.

The clots included either a deep vein clot in the leg and/or a clot that broke off and lodged in the lung (pulmonary embolus). The clots occurred between 11 and 101 days after starting protease inhibitor therapy. Clots occurred in patients taking each of the four approved protease inhibitor drugs. Two patients had a known risk factor for developing clots; one woman was taking oral contraceptive pills, while another had a slight elevation of anti-cardiolipin antibody. All six patients were able to walk and were not bed-bound or incapacitated (another risk factor for clots).

It was extremely difficult to treat four of the six patients, since a daily dose of coumadin (an anticoagulant, or blood thinning agent) was not effective; these four had recurrent blood clots. Among those patients who discontinued their protease inhibitor therapy, no recurrent clots occurred, even after stopping anticoagulant therapy.

The authors conclude that protease inhibitors increase the risk of developing a life-threatening blood clot. This reviewer is not aware of any past reports of this association. Women who are taking both a protease inhibitor and birth control pills should be aware of this possible increased risk.

• New definition of lipodystrophy syndrome is proposed.

• Lipodystrophy rate of 84% reported after 22 months of protease inhibitor therapy (including 12% severe cases).

• Abnormal blood fat levels occur in 90% in one study.
• Glucose (blood sugar) tolerance is abnormal in 23%.
• Diabetes occurs in 15%.
• Stanford researchers recommend routine fasting blood sugar tests for all patients taking protease inhibitor therapy.
• New hypothesis proposed to explain metabolic side effects of protease inhibitors.

An entire session and one late-breaker report were devoted to the topic of protease inhibitor side effects. David A. Cooper, MD, of St. Vincent’s Medical Center in Sydney, Australia, described his group’s proposal for a definition of lipodystrophy syndrome. The new definition includes peripheral wasting of the arms, legs, buttocks, and face; central fat accumulation ("protease paunch," "buffalo hump," female breast enlargement); increased blood levels of fats (cholesterol, triglycerides); and insulin resistance, sometimes with frank diabetes.

In a late-breaker oral presentation, Andrew Carr, MD, also part of the St. Vincent's team, provided an update on rates of metabolic complications due to protease inhibitors. A total of 116 people with HIV/AIDS taking one or two protease inhibitors in 1997-1998 were compared to a control group of HIV positive patients taking anti-HIV therapy without protease inhibitors, and to an HIV negative control group. Among those taking protease inhibitors, 30% had an AIDS diagnosis and 98% were men. All approved protease inhibitors and amprenavir were used by some of the patients, with indinavir being the most common. The mean duration of protease inhibitor therapy was 22 months.

Total body fat percentage was significantly lower in the protease inhibitor arm (15%) than in the control arm (24%); total fat weight was also significantly lower in the protease inhibitor arm. When comparing different body locations, the protease inhibitor group had a significantly lower fat weight in their arms, legs, and trunk. However, there was no significant difference when comparing fat in the central abdomen. There was a mean decline of 0.2 kg of fat per month for those taking protease inhibitors. For two-thirds of the protease inhibitor group, the total fat loss was more than 5% during the observation period.

Patients were asked to rate their lipodystrophy as nil, mild, moderate, or severe. The ratings were confirmed upon physical examination by the physician. Eighty-four percent rated their overall lipodystrophy as mild to severe (42% mild, 29% moderate, 12% severe); only 12% rated it as nil. Fat accumulation in the abdomen was rated mild, moderate or severe by 68%, compared to 18% who reported mild to severe fat accumulation in the upper back ("buffalo hump"). When comparing patient and physician ratings of lipodystrophy, the reported severity correlated significantly with body fat percentages on dual energy X-ray absorptiometry (DEXA) scans.

Fasting triglyceride levels, fasting insulin levels, fasting glucose, and fasting cholesterol levels were abnormally high in those taking protease inhibitors. Cholesterol levels were abnormally high (greater than 200 mg/dl) in 77% of those in the protease inhbitor group, compared to 30% of controls. Triglyceride levels were abnormally high (greater than 176 mg/dL) in 66% of those in the protease inhibitor group, compared to 23% of controls.

Glucose intolerance (a measure of abnormal sugar metabolism) was present in 23% and frank diabetes occurred in 1% of those taking protease inhibitors. An oral glucose tolerance test (another measurement of glucose metabolism) revealed diabetes in 7% and a lesser abnormality in another 15%. Abnormal glucose tests did not correlate with lipodystrophy. Due to the high rate of abnormal glucose metabolism, the authors recommend glucose tolerance tests for all patients taking protease inhibitors. In a separate presentation, researchers from Stanford University Medical Center concluded that "routine monitoring of fasting glucose is warranted in all patients taking protease inhibitors."

Cooper described his group’s hypothesis explaining many of the side effects of protease inhibitors. The hypothesis was published in the June 20, 1998 issue of The Lancet, and is sure to generate much discussion. The research team found that HIV’s protease enzyme has a 60% homology (matching elements) with two normal human proteins (CRABP-1 and LRP) that are involved in fat metabolism. They propose that protease inhibitors, in addition to inhibiting the HIV protease, may also at least partially inhibit these normal proteins. Binding CRABP-1 causes decreased retinoic acid production, which causes fat cells in the periphery of the body (arms, legs, face) to undergo apoptosis (programmed cell death), leading to the release of fat from these cells into the bloodstream. Altered retinoic acid production may also explain other side effects including dry lips and skin, hair loss, and ingrown toe nails (so-called "ectodermal dysplasia"). Binding LRP causes abnormal fat metabolism by the liver, which leads to increased blood levels of several types of fats, including cholesterol and triglycerides. The resulting elevation of blood fats leads to fat redistribution to central locations in the body, including the abdomen, waistline, base of the neck, and female breasts. The increased blood fats also cause insulin resistance, which in genetically susceptible individuals may lead to glucose intolerance and type 2 (adult onset) diabetes.

Cooper and Carr claim that the entire lipodystrophy syndrome should be reversible, assuming that the binding of CRABP-1 and LRP is not permanent. They state that the wasting in the arms, legs, and face may take longer to return to normal than other abnormalities. The researchers called upon the pharmaceutical industry to develop protease inhibitors that do not bind either CRABP-1 or LRP and to routinely check for effects of candidate drugs on human proteins.

Cooper noted that their hypothesis must be confirmed and that more research must be done in areas such as predictors of side effects, lipodystrophy in women and children, correlation of side effects with different protease inhibitors and different doses, and the benefits of treatments including diet modification, cholesterol-lowering drugs, and diabetes medications. He reminded attendees of the irony in a widely repeated quote from the International AIDS Conference in Vancouver just two years ago: "Highly active antiretroviral therapy including a protease inhibitor will make HIV infection a chronic manageable disease just like...diabetes."

Eighteen percent (21 of 118) of women taking one of the four approved protease inhibitors in a Rhode Island study noted body changes after a mean of 14 months, confirmed by bioelectrical impedance analysis (BIA) measurements.

Changes in body fat distribution included:

• increased size of abdomen ("protease paunch") in 90%
• increased breast size in 71%
• weight gain in 62%
• decreased size of arms and/or legs in 43%
• decreased size of buttocks in 38%
• cervicodorsal fat pad ("buffalo hump") in 19%
• waist circumference greater than 34.5 inches in 70%
• a body mass index (a correlation of weight to skin surface area) greater than 27.3 kilograms per square meter in 57%.

Blood fat abnormalities included:

• total cholesterol greater than 200 mg/dL in 48%
• high-density lipoprotein "good" cholesterol (HDL) less than 50 mg/dL in 56%
• low-density lipoprotein "bad" cholesterol (LDL) greater than 130 mg/dL in 48%
• triglycerides greater than 200 mg/dL in 36%.

The authors concluded that, "A change in body habitus among [HIV positive] women associated with serum lipid abnormalities is a highly prevalent long-term side effect of HAART which includes a protease inhibitor."

In an interesting poster authored by Ramon Torres, MD, from St. Vincent’s Medical Center in New York City, abnormal fat accumulation in the abdomen and at the base of the neck were partially reversed with recombinant human growth hormone (Serostim). Human growth hormone is approved for AIDS-related wasting. The dosage used was 5-6 mg, self-injected daily for 12-24 weeks. Growth hormone is extremely expensive and is not covered by all health plans.

The poster included before and after photographs of three patients showing an obvious reduction in prominent "buffalo humps" and "protease paunches." After 12 weeks of therapy, one woman had a complete resolution of her "buffalo hump" and a 50% reduction in her abdominal "paunch." "Buffalo hump" size reductions ranged from 25% to 100%. Side effects from growth hormone included carpal tunnel syndrome (leading one patient to discontinue), joint aches, increased blood pressure, and increased amylase levels.

Torres concluded that recombinant human growth hormone may be an effective treatment for protease inhibitor-induced "buffalo hump" and "protease paunch," and believes this therapy deserves further investigation.

Lipodystrophy Reversal Following Protease Inhibitor Switch

• Lipodystrophy partially reversed after three months in 33% of 21 patients after protease inhibitor was changed to nelfinavir from indinavir or saquinavir/ritonavir.
• No change in lipodystrophy occurred after three months in 57%; the condition worsened in 10%.
• Mean triglyceride levels decreased by 0.9 nmol/L, mean weight increased by 2.4 kilograms, and cholesterol levels were unchanged three months after the switch.

Diabetes Resolves After Protease Inhibitor Change

• 63% (5 of 8) of male patients’ diabetes resolved after their protease inhibitor drug was changed to nelfinavir from indinavir.
• Diabetes had developed in all eight men after starting indinavir-containing combination therapy.
• 50% (4 of 8) of the men had a family history of adult-onset (type 2) diabetes; all eight were racial/ethnic minorities.
• Six of eight men switched directly from indinavir to nelfinavir
• In three of these six, no recurrence of diabetes was seen after 5-6 months of taking nelfinavir in three of six men; the remaining three of these six still have diabetes.
• Two of eight men stopped their indinavir and their diabetes resolved within 2-6 weeks; after later adding nelfinavir, blood sugar tests remained normal during 4-5 months of follow-up.
• Indinavir may cause more insulin resistance than nelfinavir.

Gout as a Side Effect of Protease Inhibitor Therapy

• Gout was seen in two patients taking protease inhibitors.
• Increased blood levels of uric acid occurred in 17-38%.
• Patients taking protease inhibitor therapy should have regular tests of blood uric acid levels.

Swiss researchers concluded that abnormally high levels of uric acid and high blood fat levels are common among HIV positive patients taking protease inhibitors. Their cohort included 131 patients taking protease inhibitors and 25 patients taking protease-sparing anti-HIV therapy.

High levels of uric acid can lead to gout, a painful inflammation of a bone joint, usually the large toe; this occurs when uric acid crystallizes in the joint. Sometimes uric acid can crystallize in the kidney causing painful stones; rarely kidney function may be diminished. The authors imply that all patients taking a protease inhibitor should have regular blood testing for uric acid levels, probably monthly for at least three months after starting therapy.

Page last updated 6 October 1998

October 1998 Research Notes -- Part 2

About AIDS | Treatment | Prevention | Policy | About SFAF | Take Action | Donate
Español | Search | Site Map | Feedback | Email Updates | Home (sfaf.org)