icon-    folder.gif   Conference Reports for NATAP  
  XIX International AIDS Conference
July 22-27, 2012
Washington, DC
Back grey_arrow_rt.gif
Co-Morbidities Report from 14th International Workshop on Co-morbidities and Adverse Drug Reactions in HIV and XIX International AIDS Conference
  Todd Brown, MD, PhD
Associate Professor of Medicine and Epidemiology
Johns Hopkins University

Non-AIDS Co-morbidities continue to be a hot topic in HIV research. At the Comorbidity workshop (aka the Lipodystrophy Workshop) and IAS in Washington last month, multiple studies were presented which shed light on the epidemiology, pathogenesis, and treatment of various co-morbidities observed in HIV-infected patients. While there were no major breakthroughs, this report reviews the major studies in co-morbidity research at these two conferences, particularly focusing on those with clinical relevance. Studies at the Co-morbidity Workshop are referenced by "CW" followed by their abstract number and IAS abstracts are identified by their abstract numbers.

Non-HIV Medications and Co-morbidities

Medications used to treat various non-HIV related problems can contribute to the risk of non-AIDS comorbidities in HIV-infected populations. While these associations are well documented in the general population, their contribution to the risk of various co-morbidities is just beginning to be understood. Identifying these associations is important since, in many instances, these medications are more commonly used in the HIV-infected populations compared to the general population and the negative effects may be reversible with discontinuation of or substitution for the offending medication.

An excellent example of this is the use of atypical anti-psychotic medications. These medications are very effective in the treatment of thought disorders and have few extra-pyramidal side effects. Because of their relative ease of administration and efficacy, their use has expanded to other psychiatric conditions. These medications however are also associated with metabolic disturbances, including rapid weight gain, hyperglycemia, and hypertriglyceridemia. In a cross-sectional study of 2229 HIV-infected adults receiving ART enrolled in several cohorts as part of the HIV Neurobehavioral Research Program, including CHARTER, 258 participants (12%) were also receiving atypical antipsychotics (Co-Morbidity Workshop, #P07; IAS MOPE097). Compared to those not receiving these medications, atypical anti-psychotic users had a higher prevalence of diabetes mellitus and higher triglycerides, blood pressure, and BMI, after multivariable adjustment. While it is difficult to conclude from this study alone that these medications were directly responsible for these metabolic abnormalities as it has been difficult to disentangle the specific effect of these medications from effects of the underlying psychiatric disease (eg depression) and associated behaviors, this study does raise an important issue in clinical practice and clinicians treating HIV-infected patients with co-morbid psychiatric disease should be aware of these adverse metabolic effects.

Other commonly used non-HIV medications may have adverse effects on bone metabolism. Proton pump inhibitor use, for example, has been shown to be an independent risk factor for low bone mineral density and fracture in the general population, perhaps by impairing calcium absorption. In the Veteran Aging Cohort Study (VACS) (IAS MOPE087), 36% of the HIV-infected male population were receiving PPIs and its use was associated with a 80% increased risk of fracture (Hazard Ratio 1.81, 95% CI 1.48, 2.20) after multivariable adjustment. This study reminds us about the potential adverse effects of this seemingly well-tolerated (but not benign) ubiquitous medication class. We should be checking our patients' drug list to be sure that if a PPI is being taken, it is truly needed. Also in this study, protease inhibitor use, the VACS index (a score of concomitant co-morbidity), alcohol use, and cerebrovascular disease were also independently associated with fracture risk.

Biomarkers and the Risk of Co-Morbidities

The inflammation hypothesis continues to gather steam. There were several abstracts which highlighted the relationship between systemic inflammation and the risk of non-AIDS co-morbidities among HIV-infected patients. (For a broader view of HIV and Inflammation studies at IAS, the reader should check out David Shepp's excellent piece at http://natap.org)

At the Co-Morbidities Workshop, an analysis of the control arms of SMART, ESPRIT, SILCAAT (i.e those on continuous ART) was presented which investigated whether higher baseline concentrations of CRP, IL-6, and D-dimer were associated with incident non-infectious cancer (CW, #O02) and found that higher baseline concentrations of IL-6 were associated with 28% increase of incident non-infectious cancer (HR 1.28, 95% CI 1.03, 1.60) after adjustment for CD4 cell count, demographics, and the other 2 biomarkers.

Similarly, in ACTG 5202 (THLBB06), a clinical trial of ART-naïve patients randomized to efavirenz or atazanavir/ritonavir plus either abacavir/lamivudine or tenofovir/emtricitabine, the association between baseline and time-updated concentrations of systemic inflammatory markers and incident AIDS and non-AIDS events were investigated. Overall, there were 18 non-AIDS events during follow-up, including diabetes (n=6), pneumonia (n=5), cancer (n=4), and cardiovascular disease (n=3). Those with a higher baseline concentrations of hsCRP or IL-6 had a higher risk of developing a non-AIDS event, independent of CD4 cell count (hsCRP (per log ug/mL higher): HR 1.66 (1.14 to 2.43), P = 0.008; IL-6 (per log pg/mL higher): HR 1.81 (1.01 to 3.25), P = 0.047). Interestingly, in the time-updated analyses, higher concentrations of TNF alpha were associated with an increased risk of a non-AIDS event (HR 3.87 per log pg/mL higher (1.34 to 11.18), P = 0.012), suggesting that while on treatment, persistently high TNF-alpha may be important. While the results of these two studies are intriguing, it's not clear whether these relationships are causal, what the best markers are to measure, and what to make out of the apparent differences in the associations between the markers (eg. baseline associations with IL-6 vs on-treatment associations with TNF-alpha). One big question is whether strategies to reduce IL-6, TNF-alpha, or other markers of inflammation/immune activation will decrease the incidence of co-morbidities.

What about aspirin as treatment of immune activation?

There were some preliminary data presented showing that aspirin, a medication that is cheap and widely available, might be a good intervention to decrease immune activation. In this small, but intensive study (THAB0202), 25 HIV-infected individuals on ART were given ASA 81 mg over 1 week. Markers of platelet aggregation (in response to the agonists adenosine diphosphate, arachidonic acid, collagen, and epinephrine), cellular markers of immune activation (HLA-DR+/ CD38+ on CD4 cells and CD8 cells), a marker of monocyte activation (soluble CD14), and markers of systemic inflammation (CRP and IL-6) were measured pre-treatment and after 1 week of aspirin. Measures of platelet aggregation, monocyte activation, and T-cell immune activation all decreased significantly after 1 week of aspirin, but not to levels of HIV-uninfected healthy volunteers. Average concentrations of CRP, IL-6, and D-dimer were nominally decreased after 1 week of aspirin, but significance testing was not reported. All in all, this is a very interesting study and paves the way to further placebo controlled trials of aspirin in HIV-infected patients. As we continue to evaluate aspirin in further studies, it is important to keep in mind whether the decreases in the cellular markers of inflammation and platelet aggregation that were observed are of a large enough magnitude to be clinically significant. This is difficult to predict. Clearly, immune activation did decrease but did not normalize (at least compared to young healthy volunteers). This will be an important question going forward as other interventions are explored for their effect on inflammation and immune activation. One issue specific to aspirin is the potential risk of bleeding (GI and hemorrhagic stroke) which must be factored into the risk/benefit equation. There has been increased recognition of the potential adverse effects of aspirin and this has prompted revisions of primary prevention recommendations, even in patients with higher cardiovascular risk, such as those with diabetes.

Other Potentially Important Biomarkers Associated with Co-Morbidity Risk

Low CD4 cell count : Poor immune recovery in the setting of successful ART is another marker of co-morbidity risk which may be related to persistent immune activation. In the MASTER Cohort (CW O03; MOPE102), of the 1228 study participants with a CD4 count < 200 cells/mm3 prior to ART initiation who also had an undetectable viral load after ART initiation, 26% still had a CD4 cell count < 200 after 1 year on treatment. During this time, 102 new non-AIDS diagnoses were made (25 malignancies, 21 severe infections, 19 renal, 16 cardiovascular, 9 hepatic, 6 other nonfatal illnesses, and 6 deaths). In multivariable models, poor immune recovery (CD4 < 200 on treatment) was associated with a higher risk of non-AIDS diagnoses (adjusted hazard ratio [aHR] 1.96, 95% CI 1.21 to 3.19, P = 0.006). These findings highlight an important population in whom specific therapies to decrease immune activation should be directed. In addition, this population may also particularly benefit from tighter control of traditional risk factors for co-morbid disease (eg HTN, LDL cholesterol, diabetes mellitus, smoking cessation, etc for CVD).

AGE in the Skin: We all know from looking at ourselves in the mirror that the skin is another organ where aging is apparent. Using a non-invasive technology (skin autofluorescence), researchers in the AGEh Cohort in Amsterdam investigated whether the accumulation of advanced glycation end products (AGE) in the skin was associated with HIV-infection and the presence of co-morbidities. AGEs develop as a result of the non-enzymatic glycation of proteins, lipids and DNA, and are well known in the diabetes world. AGE accumulation is thought to be a major player in the pathogenesis of end-organ complications in uncontrolled diabetes. AGEs also increase with age, smoking, renal disease, and inflammation.

In this cohort of 489 HIV infected and 452 HIV-uninfected persons (average age ~52, ~85% male), HIV-infected persons had higher AGE compared to the uninfected controls across the age (age in years, that is) spectrum. In addition, higher levels of AGE were associated with an increased prevalence of co-morbidities (hypertension, cardiovascular disease, diabetes, COPD, renal impairment, cancer, osteoporosis), independent of HIV. HIV-infection was of borderline statistical significance in the model which also included AGEs. Interestingly, the magnitude of the HIV-effect on the odds of having more co-morbidities decreased only slightly after taking the AGE (in CAPS) into account. Taken together, these findings suggest that although AGE levels are higher in HIV-infected patients, they do not account for the higher burden of co-morbidities in this cohort. This argues against AGEs being in the causal pathway between HIV and co-morbid disease. Of course, this is a cross-sectional study, so it's impossible to get a sense of causation. It will be important to examine the accumulation of AGEs is prospective studies to understand its ability to predict non-AIDS co-morbidities. Also, the various co-morbidities have separate pathogeneses and may be differentially affected by AGE accumulation. Future studies should examine AGE/HIV relationships within a single disease (eg cardiovascular disease).

Cardiovascular Disease

There were few major studies regarding cardiovascular disease in HIV. One study that caught my eye examined CVD at very early stages. This was a small investigation of young HIV-infected patients (n=20) without known cardiovascular disease or symptoms (CW O10, THPE084) and HIV-uninfected controls in which a novel MR imaging technique with a cool name ("black-blood vessel wall imaging") was used to assess the arterial wall thickness in the right coronary artery. Despite their younger age (mean age 21 v 29 years) and lack of traditional CVD risk factors, HIV-infected individuals had significantly thicker coronary arteries compared to HIV-uninfected controls, suggesting very early vascular abnormalities associated with HIV infection and its treatment. Clearly, larger studies using this sensitive technique are warranted.

Cerebrovascular Disease: The Other CVD

Studies demonstrating higher risk of stroke in HIV-infected persons compared to HIV-uninfected persons continue to emerge. In a well-done, retrospective case-control study of HIV-infected and -uninfected persons admitted to a single New York hospital with an acute stroke (THAB0204), the 41 HIV-infected stroke patients were more likely to be male, non-white, and currently smokers compared to the 101 randomly selected HIV-uninfected stroke patients. Notably, the average age in the HIV-infected patients was 57.2 years vs 72.4 in the HIV-uninfected group (P = 0.001). There was a lower prevalence of hemorrhagic stroke in the HIV-infected patients, which may account for their lower stroke severity.

In another investigation of stroke in HIV-infected persons in North Carolina found a higher than expected stroke incidence compared to the HIV-uninfected ARIC Cohort (THPE092). In addition to age, hypertension and dyslipidemia, low CD4 cell count and uncontrolled viremia were independently associated with incident stroke, reinforcing the importance of HIV-related variables and echoing the findings of the MASTER Cohort (described above) which identified low CD4 cell count as a risk factor for non-AIDS co-morbidity, such as stroke.

Other Co-Morbid Conditions: Diabetes and Cognitive Function

The inter-relations among co-morbid conditions in HIV-infected patients are critical to our understanding of their pathogenesis and optimal treatment approaches. In the general population, disorders of glucose metabolism, such as diabetes and insulin resistance have been associated with poor cognitive performance. Among HIV-infected populations, these associations have not been well defined. This may be particularly important given the high prevalence of diabetes and insulin resistance in some HIV-infected populations. In the Aquitaine Cohort, 400 HIV-infected patients underwent cognitive testing as well as assessments of glucose metabolism (THAB0203). Of these, 9.5% were found to have diabetes and 8% had impaired fasting glucose (prediabetes). In a multivariable model, including demographic factors, HIV-disease related variables, CVD risk factors, and depression, diabetes was associated with worse performance on multiple cognitive tests. In addition, in the seventy-two percent who also had follow-up data after 2 years, diabetes was associated with steeper declines in cognitive function. Although it's tempting to conclude from this study that there is a causal association between glycemic control and cognitive impairment, the true test of causality will be whether interventions that decrease blood glucose will also lead to improvements or stabilization of cognitive function. Clearly, these types of clinical trials are needed.

Is Raltegravir Bad for Muscle?

Investigators from Australia followed up on reports of CK elevation from clinical trials of raltegravir by conducting a prospective study assessing CK elevations, myalgias, and myopathy in HIV-infected persons receiving (n=159) or non receiving (n=159) raltegravir (CW O016). Skeletal muscle toxicity was defined as either: (1) isolated CK elevation; (2) myalgia without motor weakness; (3) proximal myopathy on physical examination; or (4) rhabdomyolysis. While there was no difference between the 2 groups in the prevalence of CK elevation, myalgia without weakness (19% versus 3%, respectively; P<0.01); and proximal myopathy (4% versus 0%, respectively; P=0.03) were more common in the RAL-treated group. It will be interesting to see how this story evolves, whether it's confirmed in other studies and whether it's related to other integrase inhibitors. In their presentation, the investigators noted that the incidence of CK elevation in clinical trials of raltegravir was notably higher than clinical trials of statins (5-13% vs <0.5%). A separate study from Spain reported a 11% prevalence of CK elevations in RAL-treated patients (CW P20). This study did not have a non-RAL control population, but identified male sex and prior CK elevation as independent risk factors. We are likely going to be hearing more about this potential toxicity of RAL in future studies.