8th Annual Retrovirus Conference
ADVERSE EFFECTS & COMPLICATIONS
Section written by Gilbert Kaufmann, MD, with selected contributions by Harvey S. Bartnof, MD, and Jules Levin
The lipodystrophy syndrome (fat redistribution) is one of the major adverse events affecting a large proportion (20-80%) of individuals with HIV who are treated. The exact cause(s), however, are still unclear, but changes in adipogenesis (formation of fat tissue) and adipolysis (breakdown of adipose tissue) have been reported. There is much controversy about the diagnostic criteria (what constitutes the syndrome) and the true incidence (rate), as well as whether it is only one syndrome or several syndromes. Initially associated with the use of protease inhibitor (PI) drugs, subsequent studies provided evidence that the use of nucleoside (NRTI) drugs and non-nucleoside reverse transcriptase inhibitor (NNRTI) drugs may similarly be associated with this syndrome. Potential co-factors include HIV itself and affect of HIV and therapy on immune decline and reconstitution, age, race-ethnicity, gender (sex), genetics, co-infection with HCV, hormones, and diabetes/insulin resistance.
Dr. C. Goujard of the Primo Group in France presented a study that evaluated the incidence of the lipodystrophy (body changes: face, limbs, trunk, abdomen, breast) syndrome in 121 patients with acute (primary) HIV (recent) infection who have been treated with HAART (highly active antiretroviral therapy) for more than 6 months (abstract 403). Twenty-two patients (18.2%) presented with at least one clinical sign of lipodystrophy after a mean follow-up of 24 months. The proportion of patients with lipodystrophy syndrome increased to 30% at 36 months. This study confirms previous reports that lipodystrophy can already occur in very early stages of treated HIV infection and it suggests that duration of HIV infection might not be a major cause.
Dr. L. Kingsley of the University of Pittsburgh studied the prevalence (rate) of the lipodystrophy syndrome and metabolic abnormalities in the Multicenter AIDS Cohort Study (MACS), consisting entirely of men (abstract 538). A total of 868 participants underwent regular physical examinations and laboratory testing. The results showed that changes of body shape were detected in 33% of HIV negative and in 63% of HIV positive men. Various studies have found lipodystrophy rates ranging from 20%-80%. The average age of the men in this study was 47, which likely affects some of the findings in this study as increased fat accumulation and other body changes occur as men grow older. Due to the high prevalence (rate) of body changes in HIV-uninfected subjects, the investigators tried to find a better case definition of the lipodystrophy syndrome that would allow a better "discrimination" between the actual lipodystrophy syndrome and other, unspecific body changes primarily associated with increasing age. A symptom combination of moderate or severe "peripheral wasting" (thinning of arms or legs) and central adiposity (increase in abdomen or breasts, or loss in buttocks) was the best working definition. This resulted in a relatively low prevalence of 20% of patients with signs of moderate/severe lipodystrophy in the HAART-treated group, and 8% in the mono-combo NRTI treated group. But, 42% receiving HAART, and 40% on mono/combo NRTIs experienced "any" body changes compared to 10% of HIV negative men. Elevated triglycerides (>400 mg/dL) was associated with being on HAART (12%). Elevated fasting glucose was comparable whether on HAART (11%) or mono/combo NRTIs (9%). Hyperinsulinemia was found in 14% of men on HAART, 9% on mono/combo NRTIs, 0% for men on no treatment, and 7.6% of HIV negative. Low HDL (the good cholesterol) may be a hallmark of HIV-infection and not necessarily just HAART use: 0% in HIV negatives and 22%, 27%, and 29% incidence rates in HIV-infected taking no treatment, mono/combo, and HAART, respectively. Although finding a case definition for the lipodystrophy syndrome is critical to estimate the true prevalence of this syndrome, it may not be possible due to the many various contributing factors and manifestations. A further interesting finding of this study was that the prevalence of moderate to severe lipodystrophy reached a plateau after 2 years, but the incidence of "any" lipodystrophy continued to rise through the 4 year follow-up.
Dr. V. Joly of the ANRS in Paris, France presented the NOVAVIR study (abstract 539). A total of 170 patients were randomized to receive either AZT (Retrovir, zidovudine, NRTI drug) or d4T (Zerit, stavudine, NRTI drug), in combination with 3TC (Epivir, lamivudine, NRTI drug) and indinavir (Crixivan, PI drug). But this analysis only includes 96 of the 170. The results after 30 months revealed that the proportion of patients with signs of lipoatrophy (fat loss under the skin) was significantly larger and the number of clinical signs of lipoatrophy was significantly higher in the d4T arm. There were no significant differences between the two arms in the rates of central fat accumulation. The study suggests that d4T increases the incidence of lipoatrophy compared to AZT. Pretreatment with NRTI drugs (AZT, ddI, or ddC for at least 6 months) represented a potentially confounding co-factor. A strength of this study is that its randomized, but the doctors conducting skinfold measues to detect change in lipoatrophy were not blinded to which patients were taking AZT or d4T. See additional review of this study in Clinical Studies (NRTI section).
As has been reported previously, 2 small studies showed that the lipodystrophy syndrome including abdominal fat accumulation and peripheral fat loss could affect HIV positive children taking HAART in a similar way as adults (R.A. Amaya abstract 649 and G. Meneilly abstract 650). Moreover, a large proportion of these children had hypercholesterolemia (increased blood cholesterol), hypertriglyceridemia (increased blood fats) and "insulin resistance" (pre-diabetic state or diabetes).
A significantly higher rate of lipoatrophy was observed among HIV/HCV (hepatitis C virus) co-infected patients taking HAART than among HIV mono-infected patients taking HAART (abstract 654). Differences in "insulin resistance" also were reported. For more information, see this section.
Changes Associated with Changes in Blood Markers
One of the more interesting posters about lipodystrophy and lipid levels at the Retrovirus Conference was presented by Dr. Nevena Christeff of the Pasteur Institute in Paris, France (abstract 660). Dr. Christeff found that in a prospective study of 31 HIV positive men taking anti-HIV therapy (94% HAART with a PI and NRTI drugs), changes in certain blood markers were associated with worsened, improved or the development of lipodystrophy (LD, fat redistribution). Twenty HIV negative, control men were also included. The men were followed prospectively up to 24 months. At baseline, 20 of those with HIV did have LD, as determined by CT (computed tomography) scans and physical examination, while the remaining 11 HIV positive men did not.
After the follow-up interval, those with LD at baseline were classified as "unchanged" (9 of 20), "aggravated" (worsened, 5 of 20), or "improved" (6 of 20). Those without LD at baseline were classified as "evolution" to LD (2 of 11) or "absence of evolution (9 of 11). Evolution of LD was determined by physical examination by the physician, patient report and "anthropomorphic" measurements including waist-hip ratio and skin-fold thickness. The major conclusions of the study were that:
Among men with LD at baseline, worsening of LD was significantly associated with a decrease in blood "DHEA" (dihydroepiandrosterone, steroid hormone precursor), "dramatic increase in cortisol: DHEA ratio," persistently high levels of "interferon alpha," (cytokine, intercellular messenger) and persistently high levels of blood lipids (triglycerides and "VLDL" (very low density lipoprotein cholesterol). (Note cortisol is a natural "stress" hormone released from the adrenal gland on top of each kidney; cortisol measurements were all "fasting" levels.)
Among men with LD at baseline, LD improvement was significantly associated with an increase in the DHEA levels leading to a normalized cortisol: DHEA ratio, and a normalization of interferon alpha and lipid levels.
Among men without LD at baseline, LD development (only 2 patients) was associated with a decrease in DHEA, an increase in the cortisol: DHEA ratio, and increased interferon alpha and lipid levels.
Among men without LD at baseline, the absence of developing LD was associated with elevated cortisol, "highly elevated DHEA leading to a normal cortisol: DHEA ratio," and normal interferon alpha and lipid levels.
The authors concluded, This study demonstrates that clinical evolution of lipodystrophy and lipid changes associated [with] this syndrome are related to changes in both [the] cortisol: DHEA ratio and interferon-alpha concentrations [and] also confirms our previous observations suggesting that the concomitant action of cortisol, DHEA, particularly the cortisol:DHEA ratio and interferon alpha is one of the mechanisms responsible for lipid perturbations and fat mass distribution in [the] lipodystrophy syndrome.
This study raises more questions than it answers, but might represent the next step in our understanding fat redistribution and lipid alterations in HIV patients taking HAART. Most of the studies to date about LD are cross-sectional studies; however, this one is prospective. The main question that arises is why did some patients in the current study have or develop abnormal (or normal) levels of cortisol, DHEA and interferon alpha? And, are those abnormal levels a direct cause of fat redistribution or merely a marker that occurs simultaneous to (associated with) LD? The authors reported, "none of the patients tested received any complementary treatment, including glucocorticoids [steroids] or ketoconazole [Nizoral, antifungal drug]." There is an assumption that none were prescribed male sex hormones or growth hormone. But is it possible that some were using illicit steroids outside of the study? Also, DHEA is available as a "dietary supplement" in the US. Were any patients taking supplementary DHEA not known by the study authors? Would taking supplementary DHEA improve or prevent lipodystrophy? Would "stress reduction" interventions that lead to decreased cortisol levels improve or prevent LD? The authors do not address these questions and others that are raised by their findings. The association with interferon alpha is quite interesting, since this is an approved agent to treat chronic hepatitis C and B and some HIV/HCV co-infected patients have taken it.
It is noteworthy that one study at this Conference found a higher rate of LD among HIV/HCV co-infected patients taking HAART than HIV-infected patients taking HAART (see this section). Again, this study does not establish cause & affect (that DHEA imbalances causes LD). Other potentially confounding factors in the current study include: duration of HAART at study baseline; duration of LD at study baseline; adherence to therapy; CD4 counts and HIV RNA levels (that might correlate with adherence); pre-HAART levels of cortisol, DHEA, interferon alpha and lipids; types of ART in the HAART regimens; types of fat redistribution (lipoatrophy versus fat accumulation); and exercise regimens. Notwithstanding these limitations and various unanswered questions, the study was rather provocative and certainly will lead to further studies and much discussion.
Elevated Lipid Levels Associated With an Increased Risk for Coronary Heart
Based upon HIV negative populations, it must be assumed that HIV positive patients with elevated cholesterol levels are similarly at risk for CHD (coronary heart [artery] disease). For many HIV positive persons, HAART (highly active antiretroviral therapy) appears to increase lipid (blood fat) levels. Several reports at the Retrovirus Conference addressed this issue.
Daniel Klein, MD of Oakland, California reported an update about the rate of coronary heart disease (CHD) events in 4,541 HIV-infected men receiving PI drugs in the Kaiser Permanente HMO (health maintenance organization) in Northern California (abstract 655). The cases were compared to 43,118 randomly selected, age- and sex-matched control members not known to be HIV-infected. The CHD incidence was 5.2 events per 1,000 patients per year for individuals taking anti-HIV therapy without PI drug therapy and 5.8 for patients taking a protease inhibitor. The rate in HIV-uninfected men was significantly lower at 3.4. Regarding other CHD risk factors, HIV positive patients had a significantly lower rate of hypertension (high blood pressure, 13% versus 21%), a non-significantly higher rate of increased cholesterol levels (21% versus 15%) and similar rates of diabetes (6%) and smoking (29%) as control members. The study suggests that PI drug use does not increase the short-term risk for CHD, when compared to anti-HIV therapy without a PI drug. However, the higher rate of CHD events among HIV-infected persons taking ART remained unexplained. Whether other possible co-factors for CHD, including cytomegalovirus (CMV) and chlamydiae, or HIV itself might explain the difference, is unknown.
Dr. M. Mary-Krause of INSERM in Paris, France studied the course of CHD in the French Hospital HIV Database. Myocardial infarction ("heart attack") was diagnosed in 84 of 42,787 men exposed to PI drug therapy, so the risk of heart attack was low. The incidence rate of myocardial infarction increased in parallel with the time of exposure to PI drug therapy (after 18 months of therapy, up to 30 months) and was 3-fold higher among those with at least 30 months of PI drug exposure than the expected incidence in the general population of HIV-negative men. However, the statistical analysis was not adjusted for CHD co-factors. Although this study provides some preliminary evidence that the risk for CHD is increased in patients with long-term exposure to PI drug therapy, a longer follow-up period is required to estimate the full impact of long-term elevations of lipid levels on the development of CHD in this cohort of patients.
Risk for Increased Cholesterol Associated with Anti-HIV Drugs
Dr. Andre R. Miseraz of University of Basel in Switzerland reported that a genetic variation was significantly associated with an increase in total blood cholesterol after starting a protease inhibitor drug (abstract 500). A naturally occurring "polymorphism" (variation, "genotypes 11/12") of SREBP (sterol-regulatory element binding-protein)-1c was significantly associated with an increase in blood cholesterol after starting PI drug therapy, when compared to "homozygous" (the same two gene doses) "genotype 22." A total of 67 HIV positive patients were included in the study. While the results would need to be confirmed in larger studies, they suggest that testing for the altered gene one day might help to predict who will develop high cholesterol after starting a PI drug and that that information would be useful when constructing an optimal anti-HIV drug regimen. Additionally, there might be implications for designing new drugs to treat high cholesterol for HIV-positive and HIV-negative populations.
Toxicity" & "Lactic Acidemia:" New Classification System
Andrew Carr, MD of St. Vincents Hospital in Sydney, Australia presented a State-of-the-Art lecture on "mitochondrial" (energy producer of cell) toxicity (session 64, no abstract). Different nucleoside reverse transcriptase inhibitor (NRTI) drugs are thought to inhibit mitochondrial "polymerase gamma" enzyme to varying degrees, resulting in toxicity. Elevated "lactate" levels (lactic academia, increased blood acid level) may be one of the consequences of mitochondrial toxicity. He proposed the following classification of lactic acidemia: normal less than 2 millimoles per liter (mmol/L), mild increase: 2-5, moderate increase: 5-10, and severe: greater than 10 mmol/L. Subjects with mild lactic acidemia may be asymptomatic (no symptoms), whereas patients with moderate lactic acidemia may show symptoms such as fatigue, malaise, and gastrointestinal (stomach-colon) symptoms including nausea. The severe form of lactic acidemia is often acidotic (lactic acidosis syndrome, grouping of abnormalities), which is associated with a high mortality (death rate), particularly in women who are overweight. With regard to these serious consequences, Carr recommended that NRTI drug therapy should be discontinued when lactate levels increase above 5-10 mmol/L. However, he did not recommend routine testing of lactate levels. Nonetheless, clinicians and patients should be alerted by symptoms associated with this syndrome. Dr. S.M.E. Vrouenraets of the Academic Medical Center in Amsterdam, The Netherlands reported that increased blood lactate levels in persons taking NRTI drug therapy were not consistent in the same individuals, with a wide intra-patient variability (abstract 625). This Netherlands study underscores the recommendation that testing lactate levels is difficult.
Dr. J.T. Lonergan of the University of California at San Diego found that the incidence of lactic acidemia was more than 10-fold higher for any d4T (Zerit, stavudine, NRTI drug)-containing regimen compared with a regimen without d4T (abstract 624). Patients were included if he/she had an increased ALT (liver enzyme) and/or gastro-intestinal (stomach-colon) symptoms. The greatest risk occurred when ddI (Videx, didanosine, NRTI drug) or ddI/3TC (Epivir, lamivudine, NRTI drug) was combined with d4T. However, patients with lactic acidemia could be safely rechallenged (newly started) with other NRTI drugs such as abacavir (Ziagen) or ZDV (zidovudine, Retrovir) or both after discontinuing the original NRTI drugs and the acidemia had resolved.
Dr. Vrouenraets (referenced above) found a 22% prevalence (rate) of elevated blood lactate levels among 223 patients taking NRTI drug therapy and 8% among untreated patients. Consistent with the Lonergan study above, they found a significantly increased risk for d4T-containing regimens ("odds ratio" or OR of 12), and a trend towards a decreased risk for 3TC-containing regimens (OR 0.19). In a "multivariate" analysis, the other 4 NRTI drugs were not significantly associated with an increased lactate. Neither hepatitis virus co-infection (hepatitis C or B) nor increased liver enzyme (ALT) was statistically associated with increased lactate levels.
for Metabolic Complications and Lipodystrophy
Dr. J. Miller of the University of New South Wales in Sydney, Australia evaluated the safety and efficacy of gemfibrozil (Lopid) to treat elevated triglyceride (blood fats) levels (abstract 540). Gemfibrozil was well tolerated and did not appear to be associated with additional PI drug toxicity, but had only a modest effect on hypertriglyceridemia. Gemfibrozil is FDA-approved to treat high blood triglyceride levels in HIV negative persons.
Dr. E. Martinez of Hospital Clinica in Barcelona, Spain studied the impact of substituting efavirenz (Sustiva, NNRTI drug) for a PI drug in combination therapy (abstract 668). The randomized study involved 93 HIV-infected adults with lipodystrophy syndrome. At 12 months after switching, patients on efavirenz had significant decreases in measured severity scores of several types of lipodystrophy, as well as improvements of "insulin sensitivity" and HDL-(high density lipoprotein or "good") cholesterol. However, the efavirenz arm did still have a further decrease in fat under the skin (worsened lipoatrophy), but less of a decrease than the PI drug arm. Therefore, the substitution of efavirenz (after 12 months) for a PI drug appears to slow the evolution of body fat changes except for fat loss, although it does not prevent or reverse them. Both arms maintained undetectable HIV viral loads.
In contrast, Dr. D. Estrada of Clinica de San Carlos in Madrid, Spain found somewhat different results in a non-randomized study of 41 patients with lipodystrophy syndrome (abstract 671). One year after switching to an efavirenz-based from a PI drug-based regimen, there were no significant changes in insulin sensitivity. Also, waist-to-hip ratio (one measurement of abdominal obesity) did not change significantly. HIV viral load remained undetectable in all but one patient.
Yet, Dr. R. Walli of Ludwig-Maximillians University in Munich, Germany showed that switching from a PI- to an abacavir (Ziagen, NRTI)-based regimen results in an improvement of insulin sensitivity and a decrease in both total cholesterol and triglycerides in the majority of patients up to 12 months. Changes in fat redistribution were not reported. These results are consistent with results seen from previous studies.
Although some of these reports might appear somewhat promising, it appears that some body shape changes may not be easily reversed.
and "Osteoporosis" (Loss of bone mineral density)
Several cross-sectional studies evaluated the incidence of osteopenia (mild loss of bone mineral density) and osteoporosis (severe loss). Osteoporosis is common in elderly, HIV-negative women and some elderly, HIV-negative men and increases the risk of bone fractures. Several studies at the Conference reported increased rates of bone mineral loss associated with HIV infection and not necessarily due to anti-HIV therapy. For example, Dr. Ian McGowan of Gilead Sciences reported a 24% osteopenic rate and a 2% osteoporotic rate among 151 HIV positive patients who were antiretroviral-naïve (never took any treatment) (abstract 628).
Similarly, Dr. H. Knobel of Hospital del Mar in Barcelona, Spain found osteopenia in 25% of therapy-naïves, in 40% of protease inhibitor treated, in 33% of non-protease inhibitor treated and in 16% of healthy adults (abstract 629). In this study of 80 patients, the association between the use of HAART and reduced bone mineral density was not statistically significant.
Similarly, Dr. K. Chang of Yonsei University in Seoul, South Korea reported that osteoporosis and osteopenia were not more prevalent in a HAART-experienced group than in a HAART-naïve group of patients (abstract 630). Yet in this study of 109 Asians (Korean), rates of decreased bone mineral density in the lumbar spine (low back) were quite similar when comparing HIV negative controls, HIV-positive, HAART-naïve and HAART-experienced patients (18-24% range). Interestingly, in a separate report, Dr. Chang also reported that lipodystrophy was not detected in a study of 122 Koreans, including 45 HIV patients taking HAART, 32 treatment-naïve patients and 45 healthy controls (abstract 648). Lipodystrophy was measured using DEXA (dual energy x-ray absorptiometry). The length of HAART treatment was not stated in the abstract. Metabolic changes were observed in the HAART patients, including significantly increased triglycerides, decreased HDL (high density lipoprotein or "good") cholesterol and insulin resistance. The findings suggest that manifestations of HIV and/or adverse events of anti-HIV drugs are not always identical when comparing race-ethnic groups. A lower rate of lipodystrophy has been reported previously among HIV positive African-Americans taking HAART.
Dr. A. Lawal of St. Lukes-Roosevelt Hospital Center in New York City reported that the rate of osteopenia among HIV positive men in 1993 (pre-HAART era) was similar to the rate among HIV positive men taking HAART in 1998 (abstract 627). DEXA scanning was used. However, bone mineral density in both groups was significantly lower than case-matched, HIV negative, control patients. The 36 men from 1993 had HIV-related wasting and were being treated with oxandrolone (Oxandrin) steroid hormone. The 22 HAART patients from 1998 included were being treated with recombinant (manufactured) human growth hormone (Serostim) to treat abdominal obesity.
In another study, Dr. S. Arpadi of Columbia University in New York measured the total body bone mineral content in 51 HIV-infected children and 282 healthy controls (abstract LB8). Significant reductions in bone mineral density were observed among the HIV positive children, which progressively increased with age, but were independent of PI drug treatment. Only 6% of the HIV-positive children were treatment-naïve.
These studies suggest that there is no clear relationship between HAART and reduced bone mineral density. Also, there appears to be increasing information that HIV by itself might increase the risk of losing of bone mineral content. However, the cross-sectional design of these studies represents a major limitation. The preliminary results require therefore confirmation in longitudinal, prospective, well-controlled trials. Additionally, people with HIV may also have increased incidence of risk factors for osteopenia. Potential confounding cofactors that can contribute to developing osteopenia include being sedentary, cigarette smoking, nutrition (including inadequate calcium intake), current or past steroid hormones (prior exposure to steroid treatment for PCP, use of Megesterol acetate for appetite stimulation, post-menopausal hormone deficiency), excessive alcohol intake, genetics, elevated lipids, possibly lower body mass.
Necrosis" (Bone Death Due to Inadequate Blood)
Dr. J.C. Keruly of Johns Hopkins University reported 15 cases of "avascular hip necrosis" in their HIV Clinic Cohort (abstract 637). This condition may have no symptoms (diagnosed by x-ray or other imaging study) or may be painful. If required, the treatment is hipbone replacement with a metal "prosthesis" during surgery. The "incidence" rate in the current study was 47-fold higher than in the general HIV negative population. There was a significant trend in that the annual number of cases increased from 1995 through 2000 at their institution. Risk factors associated with this specific adverse event included steroid drug (prednisone, cortisone, others) use, low CD4 count (less than 200 cells per microliter) and a longer time since HIV diagnosis, but not treatment for HIV. Nearly half of the cases had never taken a PI or NNRTI drug. Similarly, Dr. D.M. Gaughan of Harvard School of Public Health reported 5 cases of "avascular necrosis" in HIV-infected children (abstract 638). The incidence rate was much higher in this cohort than in the HIV negative, general pediatric population.
Dr. D. Boxwell from the US FDA (Food and Drug Administration) reported that the combination of hydroxyurea (HU, Hydrea, Droxia) and NRTI drug(s) may increase the risk for serious hepatotoxicity (abstract 617). Hydroxyurea is an FDA-approved drug to help treat certain cancers and sickle cell anemia. Among 34 reported cases of hepatotoxicity associated with NRTI drug and HU usage, 25% were fatal. Among all of the cases, ddI (didanosine, Videx) was reported in 91%. The combination of ddI/ d4T/HU combination may further increase the risk of fatal outcome (61% rate). The authors concluded, "Patients taking HU and NRTIs, in particular ddI and d4T, should therefore be aggressively monitored for hepatotoxicity." Regular testing of liver enzymes (ALT, AST), in addition to pancreas gland enzyme (amylase, see "pancreatitis" below) would be reasonable for patients taking NRTI drugs, and particularly when combined with hydroxyurea. Liver and/or pancreas toxicity associated with NRTI drug usage also is associated with increased blood lactate levels and life-threatening lactic acidosis (see reports above in this Section). One limitation of the study is the voluntary nature of the reporting system. As a result of these findings, the FDA has supplemented the "Black Box" warnings of the drug product information (package inserts) for hydroxyurea, ddI and d4T (stavudine, Zerit). See below: Sulkowsky and Bartlett studies for more on hepatoxicity.
(Inflammation of the Pancreas Gland)
Dr. John Moore of Johns Hopkins University reported that the risk for pancreatitis (inflammation of the pancreas digestive gland) increased 4-fold in patients treated with hydroxyurea (HU, Hydrea, Droxia). Again, the large Hopkins HIV Clinic Database was used. HU is an FDA-approved drug to treat certain cancers and sickle cell anemia. The combination of hydroxyurea with other agents such as ddI and d4T has been associated with pancreatitis in the current and previous studies. The authors concluded, "Use of [hydroxyurea] and ddI should probably be discouraged."
|< All newsletters|