icon-    folder.gif   Conference Reports for NATAP  
  21st Conference on Retroviruses and
Opportunistic Infections
Boston, MA March 3 - 6, 2014
Back grey_arrow_rt.gif
CROI 2014: Inflammation in HIV.....HIV cure, comorbidities, interventions to reduce inflammation
  David H Shepp, MD
Associate Professor of Medicine
Hofstra-North Shore LIJ School of Medicine
Inflammation and HIV Pathogenesis.
Inflammation is known to drive HIV disease progression, but the precise mechanisms causing inflammation and CD4 cells depletion are not fully understood. Studies by Doitsh et al [1] shed light on the mechanism behind the inflammatory response in HIV. Using a lymphoid tissue culture system derived from tonsils or spleens of HIV-negative donors, they demonstrated that two pathways lead to death of HIV-infected cells. Productively infected CD4 cells, which constitute only about 5% of infected cells, die by apoptosis, a non-inflammatory pathway mediated by caspase-1. Abortively infected cells, which represent 95% of all infected cells, are killed by an alternate mechanism known as pyroptosis. Pyroptosis is triggered by IFI-16 (interferon-inducible protein 16), which functions as a viral DNA sensor, and mediated by caspase-3, an enzyme which activates powerful pro-inflammatory cytokines such as interleukin-1beta and interleukin-18, initiating a strong cascade of immune activation and inflammation. Inhibitors of caspase-3 may have potential as therapeutic agents to limit both CD4 depletion and the inflammatory response in HIV infection.
CROI webcast:
Inflammation may also be an important factor in formation of the latent viral reservoir. Interferon-inducible protein-10 (IP-10, CXCL10) is a ligand for CXCR3 and functions as a chemokine attracting CXCR3+ CD4+ cells. These markers are expressed at high levels on central memory T-cells, which are both targets for productive HIV infection and cells where a large proportion of the latent HIV pool resides. In a symposium on HIV cure research, Michaela Muller-Trutwin reviewed the role IP-10 levels in HIV pathogenesis [2]. IP-10 levels correlate with levels of monocyte/macrophage activation and are elevated in primary HIV infection in humans and in the SIV macaque model. In both settings, IP-10 levels correlate with rate of subsequent disease progression and appear to predict progression better than levels of viral RNA. Strong correlations between IP-10 levels and HIV or SIV DNA in CD4 lymphocytes were also demonstrated. These studies were conducted in untreated patients and animals. Future studies will need to explore the relation between IP-10 and the size of the latent reservoir and determine if therapies that reduce IP-10 levels can limit the size of the latent HIV reservoir.
CROI webcast:
Causes of Inflammation in HIV.
Many factors play a role in chronic inflammation in HIV. The immune response to HIV itself, microbial translocation, co-infections such as CMV or HCV, toxic effects of antiviral drugs and incomplete reconstitution of down-regulatory immune mechanisms all may play a role. In the general population, aging, obesity, dyslipidemia, dysglycemia and smoking are associated with elevation of inflammatory markers. These traditional risk factors may also play a role in the chronic inflammation seen in HIV-infected individuals. Krishnan et al [3] measured interleukin-6 (IL-6), soluble CD14, IP-10, soluble TNF receptor I (sTNFRI) and II, (sTFNRII) and d-dimer and assessed traditional risk factors in 315 HIV-infected patients virologically controlled on ART. In an adjusted analysis, older age correlated with all 6 markers. Smoking correlated with IL-6, sCD14, sTNFRI & II and fasting glucose correlated with IL-6 and sTNFRII. Obesity, as assessed by waist circumference, waist-hip ratio, and BMI showed correlations with various combinations of IL-6, SCD14 and sTNFRI or II. A related study measured sCD14 and d-dimer in 689 participants in the SUN study cohort and assessed smoking and alcohol use by questionnaire [4]. Smoking was strongly correlated with levels of sCD14 levels. Heavy alcohol use (5 or more drinks on one or more occasions in a month) was associated with higher d-dimer levels. These studies suggest that in addition to HIV-related mechanisms, traditional risk factors for chronic inflammation also are important in HIV-infected individuals. Many of these risk factors are potentially modifiable and should be addressed clinically whenever possible.
Inflammation and Comorbidities.
Cardiovascular disease (CVD). In the general population, certain angiographic features have been identified as correlates of increased risk of coronary plaque rupture, leading to acute coronary syndrome, MI and sudden coronary death. Last year at CROI, HIV-infected individuals were shown by CT angiography (CTA) to have more high-risk plaque morphologies than HIV-negative individuals, including positively remodeled plaque (PRP) and low-attenuation plaque (LAP). This year, Tawakol et al [5] presented a follow-up study correlating CTA plaque morphology with arterial inflammation measured by aortic 18fluorodeoxyglucose (FDG) PET scanning. Previous studies have shown that FDG is avidly taken up by metabolically active cells, including activated macrophages, and FDG uptake correlates with pathologically proven inflammation in the aorta. Aortic inflammation is used as a surrogate for coronary inflammation because FDG PET technology cannot yet directly assess coronary inflammation. Forty-one HIV-infected individuals on stable suppressive ART with subclinical coronary artery plaque by CTA underwent aortic FDG PET. CTA features were then compared among those with measurements above and below the median on FDG PET. Compared to the low uptake group, significantly more individuals in the high uptake group had at least one LAP or at least one plaque with 2 features. There was a trend to more PRP in the high reading group that did not quite achieve statistical significance. In a multivariable analysis FDG PET reading and duration of HIV infection were independently associated with the presence of LAP.
A second exploratory study used FDG PET to assess metabolic inflammatory activity in the aorta, spleen and bone marrow in a small cohort of HIV-infected individuals controlled on ART (n=16), elite controllers (n=7) or non-controllers (n=4) [6]. HIV-negative controls matched for age, gender and Framingham risk score were also studied. Compared to controls, FDG uptake was higher at all three sites in HIV-positives, and remained higher when the comparison was limited to those controlled on ART. Two-way comparisons showed FDG uptake was highly correlated between sites for both HIV-positive and -negative subjects. For HIV-positive subjects, a broad panel of plasma and cellular immune activation and inflammatory markers and an assay for integrated HIV DNA were measured and correlated with FDG PET results. Splenic and to a lesser extent bone marrow but not aortic uptake showed correlation with soluble inflammatory markers such as sCD14, sCD163 and IL-6. Levels of integrated HIV DNA showed a trend toward correlation with splenic uptake, suggesting a possible role for spleen inflammation in determining the size of the HIV reservoir.
The association between coronary artery abnormalities and inflammatory markers was reported in a cross-sectional study of 139 HIV-positive patients controlled on ART and 92 HIV-negative participants from the MACS cohort [7]. Higher levels of CD8 T-cell activation (CD38+/HLA-DR+) and monocyte activation (sCD14, sCD163) were present in the HIV-positive group. No significant difference in coronary artery abnormalities as assessed by CT angiography were found, although HIV-positives tended to have more non-calcified plaque and more stenoses of >50%. In an analysis adjusted for age, race and traditional CVD risk factors, both CD8 activation and sCD163 (but not sCD14) levels correlated with the presence of various types of coronary artery abnormalities among HIV-positive subjects only. No correlations were found in the HIV-negative group.
Together, these 3 studies suggest systemic and vascular inflammation play a prominent role in the development of CVD in HIV-infected individuals. Although CVD in the general population also involves chronic inflammation, the pathogenesis of CVD in HIV may be more dependent on inflammatory processes.
Diabetes (DM). Elevated markers of chronic inflammation are associated with risk of DM in the general population. Betene et al [8] used data from the control arms of the ESPIRIT and SMART studies (n=3,965) to study the association between IL-6 and hsCRP levels and incident DM in HIV-infected patients on ART. Increasing baseline levels of both markers were associated with increasing risk of incident DM. In an analysis adjusted for multiple covariates, each doubling of IL-6 and hsCRP conferred a 30% and 24% increase in DM risk, respectively. Comparison to studies of these markers in the general population suggested the associated risk was similar or perhaps greater in HIV.
Fraility. Another important correlate of chronic inflammation is frailty, a syndrome defined by the presence of 3 of 5 of the following features: weight loss, low physical activity, exhaustion, low grip strength and slow gait. The correlation between frailty and inflammatory markers was assessed in the ALIVE cohort, comprised of current or former injection drug users in Baltimore, 29% of whom are HIV-infected [9]. After adjustment for other factors, elevated levels of IL-6, sTNFRI and HIV-infection were significantly associated with frailty (adjusted OR 1.79,1.34,1.56, respectively). Other factors associated with frailty included, age, low education status, prescription drug abuse, being unmarried, multiple co-morbid conditions and depression. After adjustment for multiple other risk factors, including frailty, an inflammatory index score derived from IL-6 and sTNFRI was strongly correlated with mortality in both HIV-positive and HIV-negative cohort participants.
Lung Disease. COPD is another co-morbidity common in HIV, largely because of increased rates of smoking. Other factors such as immune deficiency may contribute to the severity of COPD in HIV. The role of chronic inflammation is unknown. Crothers et al [10] analyzed lung function longitudinally in 168 HIV-infected and 147 HIV-uninfected patients with COPD in the VA Cohort Study. The two groups were matched for current smoking status. IL-6, d-dimer and sCD14 were measured at baseline. FEV1 was similar at baseline but the rate of decline was faster among HIV-infected than HIV-uninfected participants (-100 ml/year vs -68 ml/year; p=0.03). Among the HIV-infected group, after adjustment for differences in age, smoking pack-years and BMI, elevations of all 3 inflammatory markers were associated with lower FEV1 at baseline and higher sCD14 was associated with more rapid decline in FEV1. These associations with inflammatory markers were not seen in the HIV-uninfected group. The findings of this study suggest COPD in HIV may have a unique pathogenesis with chronic inflammation contributing to the severity and rate of progression.
Markers to Predict Co-morbidities. Elevated levels of certain inflammatory markers are associated with an increased risk of co-morbidities in HIV, but some markers are not widely available for clinical use. The ratio of CD4 to CD8 T-cells (CD4/CD8) is measured routinely in the clinical practice. Like many inflammatory markers, it is abnormal in most untreated patients, and improves but usually fails to normalize with virologically successful ART. Mussini et al [11] examined the occurrence of non-AIDS defining clinical events (nADEs) in 3236 patients enrolled in the ICONA Study, who achieved viral suppression on ART and had a CD4/CD8 <1.0 at the time of suppression. This cohort was analyzed for improvement in CD4/CD8 over time and subsequent occurrence of nADEs. Reaching a CD4/CD8 >1.0 during follow-up was uncommon but was more likely with younger age, higher baseline CD4/CD8 and higher baseline CD4 count. In a multivariable model adjusted for age, gender, race, mode of HIV transmission, HCV co-infection, CDC C stage, and baseline HIV RNA and CD4 count, a current CD4/CD8 <0.3 was independently associated with an increased risk of nADEs or death (adjusted relative risk 1.64, p=0.002).
In another study examining CD4/CD8 in various populations of HIV-infected individuals on ART, Serrano-Villar et al [12] found higher levels of immune activation, immunosenescence markers and skewing of CD4 cells away from naive toward terminally differentiated phenotypes when the CD4/CD8 was <0.4 as compared to >1.0. They also reported significant negative correlations between CD4/CD8 and inflammatory markers including IL-6 and sCD14, which have previously been shown to correlate with mortality in HIV. Using case-control methodology, they found lower CD4/CD8 was an independent predictor of disease progression or death in two longitudinal cohorts of patients on ART. Comparing patients who initiated ART within 6 months of HIV infection or after 2 years of infection, they also found early ART initiators were more likely to achieve a CD4/CD8 >1.0 (OR 3.6, p=0.02). They concluded that a persistently low CD4/CD8 on ART, especially when the CD4 count is >500, may represent an important marker identifying increased risk of clinical disease events and death.
Interventions to Reduce Inflammation.
Statins. Statins are widely used in the general population to lower cholesterol and reduce the risk of subsequent cardiovascular (CVD) disease. Because dyslipidemia is common in HIV, statins are also often used in this population for the same indication. However, the effect of statins on CVD risk goes beyond lipid lowering. Statins have anti-inflammatory effects that contribute to CVD reduction, although the precise mechanism is not certain. Since both CVD and chronic inflammation are increased in HIV, statins are a particularly attractive candidate drug class to reduce inflammation in this population. They have the additional benefit of wide availability and a very well known safety profile during long-term administration.
Funderburg et al [13] reported 48 week results of a randomized trial comparing rosuvastatin 10 mg daily and placebo. Eligible participants (n=147) were on stable ART with elevated T-cell activation or hsCRP and LDL cholesterol <130 mg/dL at baseline. A large panel of immune activation and inflammatory markers were studied. Most showed declines from baseline in both groups. Declines were significantly greater in the rosuvastatin group for the monocyte activation markers sCD14 and CD16+/tissue factor+/CD14dim monocytes, IP-10, the vascular inflammation marker Lp-PLA2 and various cellular markers of CD4 and CD8 activation. Although declines in hsCRP have been seen in many studies of statins in HIV-negative populations, hsCRP did not decline in this study. In companion abstracts, rosuvastatin was reported to decrease oxidized LDL [14], a reactive form that is thought to mediate vascular damage and trigger vascular inflammation, and cystatin-c a plasma marker used to estimate glomerular filtration but also increased in inflammatory states [15].
ART. Combination antiretroviral therapy (ART) has been shown to reduce many but not all measures of T-cell and monocyte activation or other components of the inflammatory cascade. Most patients treated in the chronic phase of HIV have improvement but not to levels seen in health HIV-negative individuals. Schuetz et al [16] presented studies of initiation of ART in patients with very early acute HIV infection; 14 with Fiebig I (HIV RNA only present) and 23 with Fiebig III (HIV RNA plus antibody by EIA). When compared to HIV-negative controls, pre-ART sigmoid colon biopsies showed preservation of Th17 cell number and function (assessed by cytokine expression) in Fiebig I but not in Fiebig III. Activated CD8 T-cells were elevated in sigmoid biopsies of Fiebig I patients, and higher still in Fiebig III patients, with changes in peripheral blood paralleling those seen in rectal biopsies. ART initiated within 2 days of diagnosis preserved Th17 cell number and function and normalized activated T-cells at 6 and 24 months in Fiebig I. Fiebig III patients had improvement Th17 cell number and immune activation, but did not normalize, while Th17 cell cytokine expression did not improve.
The effects of ART on inflammatory markers are generally thought to be due to suppression of viral replication and immune reconstitution rather than direct effects on the immune system. However, different antiretrovirals may have different effects on inflammation. Castillo-Mancilla et al [17] reported immunologic effects of tenofovir DF/emtricitabine given for 30 days to HIV-negatives individuals participating in a PrEP pharmacokinetic study. Markers of T-cell activation (HLA-DR+/CD38+ CD8 T-cells) and the monocyte activation marker sCD14 were significantly decreased at the end of treatment and remained decreased off treatment 30 days later. However, multiple other inflammatory markers were not altered.
Previous studies have suggested that patients switching from a boosted protease inhibitor (PI/r)-based to a raltegravir (RAL)-based ART regimen experience improvements in a variety of inflammatory parameters, changes not seen in those remaining on PI/r. It is unclear if these improvements can be attributed to better virologic control, a pro-inflammatory effect of PI/r or an anti-inflammatory effect of RAL. Massanella et al [18] measured multiple cytokines in 111 patients in two raltegravir intensification trials conducted in patients already virologically controlled on NNRTI or PI/r-based regimens. Those on PI/r-based regimens generally had higher baseline inflammatory marker levels than those on NNRTIs, but baseline inflammatory marker levels did not correlate with residual viral replication as assessed by 2-LTR circles. RAL intensification for 48 weeks did not produce significant reduction in any of the numerous markers studied except d-dimer, which decreased only in the subgroup receiving PI/r. This study suggests that improved control of viral replication or direct anti-inflammatory effects of RAL probably do not explain the improvement in inflammation seen in PI/r to RAL switches. Instead, use of PI/r may be associated with inflammatory effects.
Probiotics. CD4 T-cell populations are rapidly depleted in gut-associated lymphoid tissue during HIV infection. Especially hard hit are Th17 cells, a subset of CD4 cells that produce IL-17 and play an important role in host defense against microbial invasion in the gut. Loss of Th17 function leads to microbial translocation, an important stimulus for chronic immune activation and inflammation in acute and chronic HIV infection. Interventions to help restore Th17 cell number and function, or modulate the pathogenic potential of gut microbes could lead to reduced microbial translocation and resultant inflammation.
In animal model, Ortiz et al [19] studied the effects of combination therapy with 3 ARVs, interleukin-21(IL-21), and probiotics on immune reconstitution in the gut. Eleven pigtail macaques were treated with triple ART starting 98 days after infection with SIV. Six additionally received a daily probiotic supplement containing lactobacillus, bifidobacterium and streptococcus species and weekly doses of IL-21, a growth factor for Th17 cells. Rectal and jejunal biopsies from multiple time points up to 1 year after infection showed significant increases in CD4 cells, Th17 cells and polyfunctional Th17 cells only in those animals treated with probiotics and IL-21. IL-21 treatment did not increase levels of CD4 T-cell activation. Clinical illness was not observed in any animals treated with probiotics and IL-21, while several animals given ART alone had clinical illness related to SIV.
A small human trial of probiotics was also reported. This study randomized 30 patients virologically controlled on ART to 8 weeks of a preparation containing lactobacillus and bifidobacteruim species or control (placebo or no treatment) [20]. In recipients of probiotics, d-dimer and CRP were significantly reduced and IL-6 reduction trended toward significance, while those in the control arm showed no significant changes. However, other markers were not affected, including LPS and sCD14, suggesting reduced microbial translocation did not occur.
Exercise. Moderate intensity exercise appeared beneficial in a study of 49 sedentary, ART-treated patients [21]. Participants were enrolled in an exercise program that included one hour of brisk walking with or without 30 minutes of circuit training exercise 3 times weekly for 12 weeks. In a subset of 25 individuals who completed the program and had inflammatory marker data available, d-dimer, IL-6, hsCRP, IL-18, myostatin, and CD4 and CD8 activation markers (HLA-DR+, CD38+) all declined significantly, while sCD14 did not. Additional benefits included significant declines in BMI, waist circumference, total and LDL cholesterol. Although this intervention appeared to provide broad reductions in inflammatory markers for those completing the program, 14/49 (29%) either dropped out or had a low participation rate. Also, it was not clear how much reduction was attributable to walking alone vs. walking plus circuit training.
Aprepitant. Neurokinin-1 receptors (NK-1R) are found in cells of the nervous and immune system. Substance p is a ligand for NK-1R involved in mediating pain and nausea. It may also play a role in inflammatory responses. Aprepitant is an NK-1R antagonist approved as an antiemetic. In vitro, it reduces CCR5-expression on CD4 T-cells with a resulting antiviral effect. To evaluate the effects of aprepitant, a pilot study randomized 18 HIV-infected individuals not on ART to 14 days of aprepitant 375 mg daily or placebo [22]. In the aprepitant arm, significant declines in sCD163 and CD4 T-cells bearing the exhaustion marker PD-1 were seen. However, plasma HIV RNA, activated CD4 and CD8 T-cells, and CD4 cells expressing CCR5 did not change. Aprepitant was well tolerated but LDL and HDL cholesterol increased significantly. Studies of the anti-inflammatory effects of aprepitant in HIV-infected individuals on ART may be warranted.
Other Interventions. Several pilot studies of candidate treatments to modulate inflammation reported negative results. A single arm study of 8 weeks of sevelamer, a phosphate binding agent that also can bind LPS, failed to reduce levels of LPS, sCD14, sCD163, IP-10, or T-cell activation markers in 40 HIV-infected individuals not on ART [23]. In a randomized trial, rifaxamin, a non-absorbable antibiotic that can decrease LPS levels in patients with hepatic encephalopathy was given for 4 weeks to 43 patients virologically controlled on stable ART but with CD4 counts <350/mm3 [24]. Compared to 22 untreated controls, there was a very small difference in T-cell activation markers but no effect on LPS, sCD14 or CD4 counts. A randomized, placebo-controlled, cross-over trial conducted of mesalamine, a salicylate derivative used in inflammatory bowel disease, was conducted in 33 participants on stable ART but with CD4 counts <350/mm3 [25]. No significant differences in activated T-cells in peripheral blood or rectal biopsies, sCD14, d-dimer or IL-6 were seen.
1. Doitsh G, Galloway N, Geng X, et al. Pyroptosis Drives Both CD4 T Cell Death and Chronic Inflammation in HIV-Infected Lymphoid Tissues. Abstract 76.
2. Mller-Trutwin M. Inflammation as an Obstacle for Remission: Lessons Learned From Non-Human Primate Models. Abstract 55.
3. Krishnan S, Bosch RJ, Rodriguez B, et al. Correlates of Inflammatory Markers After One Year of Suppressive Antiretroviral Treatment (ART). Abstract 757.
4. Cioe PA, Baker J, John Hammer J, et al. Soluble CD14 and D-Dimer Are Associated With Smoking and Heavy Alcohol Use in HIV-Infected Adults. Abstract 732.
5. Tawakol A, Zanni MV, Lo J, et al. Increased Arterial Inflammation Relates to High-Risk Coronary Plaque Morphology in HIV+ Patients. Abstract 130.
6. Hsue PY, MacNabb M, Kaplan R, et al. Arterial Inflammation in HIV as Measured by FDG-PET/CT Is Associated With Splenic Activity. Abstract 131.
7. Daar ES, Post WS, Darilay AT, et al. Monocyte But Not Cellular Activation Is Associated With Coronary Atherosclerosis in the MACS. Abstract 730.
8. Btn C, Dooko A, Neuhaus J, et al. IL-6, hsCRP, and the Development of Type 2 Diabetes Among HIV Positive Patients Taking ART. Abstract 768.
9. Piggot DA, Varadhan R, Mehta SH, et al. Frailty, Inflammation and Mortality Among Aging HIV-Infected and At-Risk Injection Drug Users. Abstract 762.
10. Crothers K, Rodriguez CV, Wongrakool C, et al. Association of HIV Infection and Immune Activation With Decline in Lung Function. Abstract 774.
11. Mussini C, Lorenzini P, Cozzi-Lepri A, et al. Incidence of CD4/CD8 Ratio Normalization and Its Role in the Onset of Non-AIDS-Related Events. Abstract 753.
12. Serrano-Villar S, Sainz T, Lee SA, et al. A Low CD4/CD8 Ratio During Effective ART Predicts Immunosenescence and Morbidity/Mortality. Abstract 242.
13. Funderburg N, Clagett B, Jiang Y, et al. Rosuvastatin Reduces Immune Activation and Inflammation in Treated HIV Infection. Abstract 335.
14. Hileman CO, Semba RD, Turner R, et al. Rosuvastatin Lowers Oxidative LDL in HIV-Infected Persons On Antiretroviral Therapy: SATURN-HIV. Abstract 750.
15. Longenecker CT, Hileman Co Storer NJ, et al. Rosuvastatin Lowers Cystatin C in HIV-Infected Subjects On Antiretroviral Therapy: SATURN-HIV. Abstract 743.
16. Schuetz A, Phuang-Ngern Y, Rerknimitr R, et al.Early ART Initiation Prevents Disruption of the Mucosal Barrier and Subsequent T-Cell Activation. Abstract 77.
17. Castillo-Mancilla JR, Meditz A, Seifert S, et al. Reduced Immune Activation During Tenofovir-Emtricitabine Therapy in HIV-Negative Individuals. Abstract 340.
18. Massanella M, Llibre JM, Marfil S, et al. Effect of Raltegravir Intensification in the Cytokine Profile of Treated HIV+ Individuals. Abstract 300.
19. Ortiz AM, Klase ZA, Carmack KM, et al. Probiotic and IL-21 Treatment Promotes Th17 Cell Recovery in ARV-Treatment of Pigtail Macaques. Abstract 83.
20. Stiksrud B, Nowak P, Kvale D, et al. Decreased Levels of D-Dimer After Probiotic Supplementation in Patients Receiving ART. Abstract 342.
21. Longo V, Bonato M, Bossolasco S, et al. Brisk Walking Improves Inflammatory Markers in cART-Treated Patients. Abstract 763.
22. Tebas P Spitson S, Barrett J et al. Reduction of sCD163, SP and PD1+ Th Cell Levels in a Phase 1B Trial of the NK1R Antagonist Aprepitant. Abstract 338.
23. Sandler NG, Zhang X, Bosch RJ, et al. Sevelamer Does Not Decrease Plasma LPS or sCD14 But Does Decrease Soluble Tissue Factor and LDL. Abstract 337.
24. Tenorio AR, Wilson CC, Chan ES, et al. Rifaximin Has Marginal Impact On Immune Activation in Immune Non-Responders To ART - ACTG 5286. Abstract 339.
25. Somsouk M, Dunham RM, Cohen M, et al. Mesalamine To Reduce Immune Activation During HIV Infection: A Randomized Controlled Trial. Abstract 341.