CROI 2017: Better Living with HIV Infection
by David Alain Wohl, MD Professor, Division of Infectious Diseases, The University of North Carolina at Chapel Hill Director, North Carolina AIDS Training and Education Center Site Leader, The University of North Carolina Chapel Hill AIDS Clinical Research Site
Each medical conference has a vibe. The data, the venue, and even the host city, can shape the mood of the attendees as they crisscross on the escalators and queue at the food courts, toting their brightly-colored conference backpacks. CROI 2017 was like the Seattle weather that week in February – partly sunny with cloud-streaked blue skies but the occasional day of rain. With cheekily named sessions showcasing generally positive results, there was a collective nod of approval as we continue to close gaps, expand treatment, and identify new targets to mess with. At the same time, there were gloomy footnotes pointing to seemingly intractable challenges, like getting ART and PrEP to all in need and addressing high rates of co-morbidities, especially among our aging HIV-infected patients.
CROI, above all else, offers guidance on how to optimize well-being for those living with HIV infection, whether through avoidance of a particular antiretroviral or by adopting an intervention to prevent a comorbidity. At this year’s conference, there were no findings to suggest HIV+ people or their providers do anything radically different. There were the expected echoes of prior messages regarding the higher risk of frailty and its components among those who are HIV-infected, the dangers of smoking and substance use, as well as not wholly surprising data regarding the effects of protease inhibitors on cardiovascular disease (CVD) risk. But, the together, the data form a booming chorus, extolling us to move from tabulating the acronymic conditions that can pile-on to make long-term HIV survival miserable, to tacking them.
As suggested by the studies described below, approaches to prevent and treat co-morbid conditions that disproportionately hit people living with HIV are not hidden and undiscovered. It is not for the want of some novel (read: expensive) anti-inflammatory that HIV-positive people suffer a stroke or get lung cancer. Tools that can dramatically reduce co-morbidities abound. The problem is that we as HIV providers have yet to fully grasp them.
CVD Prevention Model
There are studies that sound alarms about a problem, and others that provide guidance on how to actually fix the problem. A nice study modeling the impact of several interventions to blunt the rise in CVD among people living with HIV conducted by a Dutch team is one of the latter (1). The interventions explored in this simulation included the very early start of ART, avoidance of protease inhibitors (see below) and abacavir, smoking cessation (see below), and intensified treatment of hypertension and dyslipidemia. Complete (100%) and partial (50%) uptake of each individual intervention was explored individually in the model, which used data collected in the ATHENA Cohort to populate the simulated patient population.
These simulations predicted all the selected interventions to have a positive effect in reducing CVD from now to 2030, but to very different extents. Earlier ART had the least impact, and cardio-friendly ART was only slightly better. In contrast, the traditional and time-tested interventions of smoking cessation and optimization of blood pressure and lipids were predicted to prevent many more CVD events, even if only half-heartedly embraced.
The Bottom Line: Although modeling studies are inherently imperfect, this complex exercise supports other data (and some common sense) regarding CVD prevention in HIV-positive people. We can fuss over ART regimen composition, but it is smoking, hypertension, lipids, glucose control and other routine health maintenance components that are the bigger tickets worth punching to prevent CVD and almost certainly other co-morbid disease.
So, the good news is that the tools we can use to beat back the specter of CVD are widely available and supported by decades of implementation research. The bad news is that we have not been as successful as need to be to apply these CVD prevention interventions to the management of HIV. This is highlighted by an analysis of data from the NA-ACCORD cohort study that found a persistent, albeit narrowing, gap between guideline indications for statin therapy and the prescription of a statin over time (2). In 2013, 53% of HIV-positive patients at NA-ACCORD clinics for whom a statin was indicated had not been prescribed this medication.
The magic bullet for HIV aging may not be some novel anti-inflammatory molecule but medications that can be found on the Walmart $4 list. Identifying effective ways to incorporate health maintaining interventions into our HIV clinics and achieve patient acceptance and adherence to these, sometimes unpopular, measures, may go a long way to addressing the age-related problems that HIV-infected persons disproportionately endure – and be a more efficient way to spend HIV research dollars.
Quantifying the Risk of Cancer after Smoking Cessation
It is a non-alternative fact (i.e., a plain old fact) that smoking causes cancer and that the risk of malignancy drops over time with smoking cessation. However, the extent that cancer risk can be expected to decline after quitting is not clear among those with HIV-infection, who appear to be more susceptible to the carcinogenic effects of cigarette smoking than those uninfected with HIV. To get at this, D:A:D investigators looked at cancers developing in those who were cancer free when entering the cohort study, and then explored associations with smoking history (3). Of 1,980 incident cancers, 242 were lung cancer, and 487 smoking-related non-lung cancers. Smoking history was categorized as never smoker, current smoker, ex-smoker at baseline, and ex-smoker during the study follow-up. At baseline, only 31% had never smoked. As expected, smokers had higher risks for lung and the other related cancers compared to never smokers. After cessation of smoking, the overall risk of cancer, as well as the risk of non-lung smoking-related cancers, dropped over the next 5+ years; however, the risk of lung cancer remained largely unchanged after quitting.
The Bottom Line: These are concerning data in that they suggest a durable effect of smoking on cancer risk that is not appreciably ameliorated by smoking cessation. This finding is in contrast to studies in the HIV-uninfected and remains unexplained. Although this is a large cohort of HIV-positive people, the number of lung cancers was small at 242 and relatively few ex-smokers had been followed long term. Therefore, with more time, additional data will be available to confirm or adjust this finding.
Meanwhile, as abundantly suggested, getting HIV-infected people to stop smoking is second only to ART as a health priority. The message from this study must not be that stopping smoking will not reduce the risk of cancer, but that it will, here it did for both non-lung smoking cancers and cancers overall.
Knowing Frailty when you see it
Think frail and conjured is the image of a stooped elder, a strong breeze away from a fractured hip and subdural hematoma. However, frailty is more complicated and is based in a theoretical framework to better understand and predict the risk for morbidity, hospitalization, and death by taking into account loss of body mass, strength, endurance, gait speed, and physical activity.
A number of studies have convinced us that the frailty phenotype is more common among HIV-infected people, although as is the case for most all age-related complications the exact magnitude of any contribution by the virus and/or ART to this risk remains murky. At CROI, there were additional reports regarding frailty, its prevalence and determinants, that underscore the need for HIV providers to appreciate the forest of this concept, more than the trees of conditions such as CVD and osteoporosis that can contribute to it:
- An analysis of 954 patients age 40 years and older (15% greater than age 60), participating in ACTG ART treatment trials and regularly assessed for fragility and neurocognitive impairment (NCI) found that the risk of falls, worsening disability, and death were highest in those with frailty compared to those with NCI alone (4). The combination of NCI and frailty was the riskiest. Of the select adverse events, falls were the most common. The authors point is that attention to frailty may be a more appropriate target for intervention than NCI.
- That the road to frailty can be a two-way street was made clear by an analysis from the ALIVE Cohort in Baltimore (5). This study enrolls and follows HIV-infected and –uninfected participants with a history of injection drug use. A strength of this cohort is that the uninfected participants suffer from many of the same conditions and environmental stresses experienced by HIV-infected people and that can confound other studies where the controls are less well-matched. Looking at the “frailty transitions” of over 1,300 ALIVE participants (a third HIV+), decreased frailty progression and increased frailty recovery were both associated with major sociodemographic characteristics including education attainment and employment. Having fewer co-morbidities, including depression, was also associated with the frailty outcomes. After adjustment for these factors, HIV infection was still associated with risk of frailty progression and lower likelihood of frailty recovery. In the fully adjusted models, HIV virologic suppression, elevated CD4 nadir (>500/mm3) and absence of a prior AIDS diagnosis were all significantly associated with reduced frailty progression and improved frailty recovery. An inflammation index based on blood IL-6 and sTNFR1 levels and applied to a subset of participants found an association between frailty progression and recovery with the inflammation index in the expected directions. This study nicely highlights not just the mutable disease-specific targets for intervention to prevent and recover from frailty, such as early and successful HIV treatment and co-morbidity treatment and prevention, but also the environmental factors that contribute to poor health. These were quite potent and included depression, education, and employment – structural issues that are not only the fodder for headlines but also have direct effects on our health.
The Bottom Line: HIV care providers need to become more familiar with frailty. There have been many CME programs on aging and inflammation, and shifting attention to what may be an overarching issue of frailty and pre-frailty is warranted given the prevalence of this syndrome and its impact on health. Just as we have seen cardiologists, neurologists, and nephrologists specialize in relevant HIV-related conditions, it is time we reached out to geriatricians to do likewise. A handful of HIV clinics have already adopted a geriatrics model and more need to follow. For example, falls risk assessments, which I suspect no card-carrying HIV provider knows how to do properly, can identify those who are at greatest risk and most likely to benefit from intervention. Those with peripheral neuropathy, which enhances falls risk, deserve particular attention. These data suggest we may help our patients more by assessing for frailty than many of the things we do such as listening to their chest with a stethoscope.
The Die is Cast at the START
The START trial famously demonstrated that earlier ART administration prevented AIDS and serious non-AIDS events. In addition, this study, like others, found an association between markers of inflammation and coagulation and these events. Expanding on this latter work, the team examined baseline (i.e., pre-ART) inflammation markers (IL-6, hsCRP, serum amyloid A [SAA], sICAM, sVCAM), an immune activation marker (IL-27), and a coagulation marker (D-dimer) and clinical outcomes among the 2,124 participants in the immediate ART group and the 2,175 participants in the deferred ART group (6).
After adjustment for age, gender, treatment arm, and region, higher levels of D-dimer and IL-6 were associated with risk for both AIDS and serious non-AIDS events. In addition, SAA and sICAM were associated with AIDS. AIDS events were driven by TB disease and non-AIDS events by CVD. IL-6 was associated with CVD risk.
The Bottom Line: Having elevated markers of inflammation and coagulation can’t be a good thing and once again we see that to be the case. Among those entering the START trial, those harboring higher levels of IL-6 and D-dimer had a heightened risk of AIDS and serious non-AIDS events. This makes sense and points to these markers being an indicator of a brewing badness that may not be clinically evident (and may not even be HIV-related). Although there was some adjustment performed it is not clear what role pre-treatment CD4 cell count played. The median entry CD4 cell count in this study was well over 600/mm3, but there was a range at baseline and a count of 525/mm3 is not the same as 925/mm3. Other factors could also contribute to having higher levels of these markers such as duration of HIV infection and an assortment of co-morbid conditions. The important message from this analysis is less that relative elevation of these markers predicted poor outcomes, and more that we need to figure out why they were elevated.
Protease Inhibitor Pluses and Minuses
The D:A:D Cohort Study first reported an association between protease inhibitor-containing antiretroviral therapy (ART) and CVD well over a decade ago. At the time, protease inhibitors were widely used to treat HIV infection, allowing for investigation of earlier drugs in this antiretroviral class for links to CVD incidence. As more patients have received darunavir and atazanavir, still the newest kids on the protease inhibitor block, they can now also be similarly scrutinized. In doing so, the D:A:D team found a significant but small association between darunavir and CVD (~59% increased risk for every 5 years of exposure) – on par with that seen with the older protease inhibitors (7). However, such an association was not seen with atazanavir. Interesting, the darunavir-CVD association did not seem to be explained by dyslipidemia. The study did not collect data on dosing and some of the darunavir-receiving patients could have been taking the higher twice daily dosing.
A clue as to why atazanavir may be an outlier when it comes to protease inhibitors and risk of CVD can be found in by separate study from the Veterans Aging Cohort Study (VACS), also presented at CROI. This analysis examined the relationship between total bilirubin levels in the blood and the risk of CVD, as determined by diagnosis codes, in tens of thousands of HIV-infected and -uninfected patients (8) (it is known that those with Gilbert’s Syndrome, a common genetic disorder in which unconjugated bilirubin levels are elevated, have a lower risk of CVD). A clear bilirubin-CVD risk relationship was seen with a decline in CVD (and heart failure and acute myocardial infarction) risk at total bilirubin levels above 0.6 mg/dL; this was the case for HIV-positive and –negative patients. There were too few CVD events among those on atazanavir to tease out whether this protease inhibitor was in any way protective.
The Bottom Line: Fewer people living with HIV in the US are being treated with protease inhibitors now that once-a-day integrase inhibitors have popped on to the scene. Therefore, the impact of the D:A:D finding of an association between CVD and cumulative use of darunavir and the failure to find a similar link with exposure to atazanavir, is diminished compared to a few years ago. That said, some patients do need to be on protease inhibitors for one reason or another and these data may influence selection of their therapy. The observation that darunavir acts like older protease inhibitors is less surprising than the observation that atazanavir doesn’t. Although the D:A:D analysis did not find bilirubin to be explanatory, it is hard to ignore the inverse relationship between bilirubin and CVD seen in the VACS cohort and previously reported data from the AIDS Clinical Trials Group (ACTG) and others demonstrating positive changes in inflammatory markers and carotid intimal thickness with atazanavir compared to darunavir.
It is important to recognize that of the 35,711 patients included in the D:A:D analysis, only 1,157 (3.2%) had a CVD event over a median 7 years of follow-up. Reassuringly, the absolute number of patients on darunavir who experienced a CVD event was low, especially among those on the drug longer term (18 patients with 5-6 years of exposure and 27 patients with >6 years of exposure). Therefore, very, very few people on the protease inhibitor experienced CVD.
Certainly, these findings may tip the balance a smidge back in favor of atazanavir but there remain important differences between these two effective protease inhibitors, including side effect profiles and drug-drug interactions, that will likely lead to little change in provider (and guideline) preferences. As described above, more cardiovascular protective juice can be had from a different sort of squeeze.
Not every co-morbidity that strikes those living with HIV infection can be prevented or reversed with good primary care. Brains, kidneys, and arteries damaged by the long-term immunosuppression of CD4 cell count depletion, cannot always be fixed. However, the preponderance of data point out where gains can be made.
HIV medicine has transitioned from an oncological mode created to manage a progressive and complicated fatal disease to a model that more closely resembles a diabetes clinic, where lab metrics are used to plot a long-term course that avoids complications. For our patients to not only survive but thrive we need a new way to support them that recognizes their outsized risk for physical and psychological co-morbidities and the entrenched toxic social and emotional challenges that feed these conditions and, also, make addressing them difficult. The next leap forward in HIV management could be the development of a comprehensive approach to co-morbidity detection and intervention – one that can be readily implemented across HIV clinics, be they in Birmingham or Boston. Designing and implementing such a program, with input from experts in prevention and geriatric medicine, can be yet another example of HIV care leading the way.
- van Zoest R, Smit M, Nichols, B, et al. Cardiovascular prevention policy in HIV: Recommends from a modelling study. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 129
- Althoff KN, Horberg MA, Eron JJ, et al. The large gap between statin eligibility and prescription among HIV+ in North America. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 619.
- Shepherd L, Ryom L, Petoumenos K, et al. Cessation of cigarette smoking and the impact on cancer incidence in the D:A:D study. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 131.
- Erlandson K, Abdo M, Robertson K, et al. Frailty has a stronger association than neurocognitive impairment with poor outcomes. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 665
- Piggott DA, Bandeen-Roche K, Mehta SH, et al. Frailty progression and recovery among persons aging with HIV and substance use. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 133.
- Baker J, Sharma S, Grund B, for the INSIGHT START Study Group. Association of inflammation and coagulation with clinical risk in the START trial. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 623.
- Ryom L, Lundgren JD, El-Sadr WM, et al. Association between cardiovascular disease and contemporarily used protease inhibitors. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 128LB.
- Marconi VC, So-Armah K, Tate J. Hyperbilirubinemia prevents cardiovascular disease for HIV+ and HIV- individuals. Conference on Retroviruses and Opportunistic Infections (CROI), February 13-16, 2017, Seattle. Abstract 127