Bayesian tools for zero counts in compositional data

The log-ratio methodology makes available powerful tools for analyzing compositionaldata. Nevertheless, the use of this methodology is only possible for those data setswithout null values. Consequently, in those data sets where the zeros are present, aprevious treatment becomes necessary. Last advances in the treatment of compositionalzeros have been centered especially in the zeros of structural nature and in the roundedzeros. These tools do not contemplate the particular case of count compositional datasets with null values. In this work we deal with \count zeros" and we introduce atreatment based on a mixed Bayesian-multiplicative estimation. We use the Dirichletprobability distribution as a prior and we estimate the posterior probabilities. Then weapply a multiplicative modi¯cation for the non-zero values. We present a case studywhere this new methodology is applied.Key words: count data, multiplicative replacement, composition, log-ratio analysis

AIDS defining opportunistic infections in patients with high CD4 counts in the combination antiretroviral therapy (cART) era: things ain't what they used to be.

INTRODUCTION: According to reports from observational databases, classic AIDS-defining opportunistic infections (ADOIs) occur in patients with CD4 counts above 500/µL on and off cART. Adjudication of these events is usually not performed. However, ADOIs are often used as endpoints, for example, in analyses on when to start cART. MATERIALS AND METHODS: In the database, Swiss HIV Cohort Study (SHCS) database, we identified 91 cases of ADOIs that occurred from 1996 onwards in patients with the nearest CD4 count >500/µL. Cases of tuberculosis and recurrent bacterial pneumonia were excluded as they also occur in non-immunocompromised patients. Chart review was performed in 82 cases, and in 50 cases we identified CD4 counts within six months before until one month after ADOI and had chart review material to allow an in-depth review. In these 50 cases, we assessed whether (1) the ADOI fulfilled the SHCS diagnostic criteria (www.shcs.ch), and (2) HIV infection with CD4 >500/µL was the main immune-compromising condition to cause the ADOI. Adjudication of cases was done by two experienced clinicians who had to agree on the interpretation. RESULTS: More than 13,000 participants were followed in SHCS in the period of interest. Twenty-four (48%) of the chart-reviewed 50 patients with ADOI and CD4 >500/µL had an HIV RNA <400 copies/mL at the time of ADOI. In the 50 cases, candida oesophagitis was the most frequent ADOI in 30 patients (60%) followed by pneumocystis pneumonia and chronic ulcerative HSV disease (Table 1). Overall chronic HIV infection with a CD4 count >500/µL was the likely explanation for the ADOI in only seven cases (14%). Other reasons (Table 1) were ADOIs occurring during primary HIV infection in 5 (10%) cases, unmasking IRIS in 1 (2%) case, chronic HIV infection with CD4 counts <500/µL near the ADOI in 13 (26%) cases, diagnosis not according to SHCS diagnostic criteria in 7 (14%) cases and most importantly other additional immune-compromising conditions such as immunosuppressive drugs in 14 (34%). CONCLUSIONS: In patients with CD4 counts >500/ µL, chronic HIV infection is the cause of ADOIs in only a minority of cases. Other immuno-compromising conditions are more likely explanations in one-third of the patients, especially in cases of candida oesophagitis. ADOIs in HIV patients with high CD4 counts should be used as endpoints only with much caution in studies based on observational databases.

Monotherapy with darunavir/ritonavir is effective and safe in clinical practice.

INTRODUCTION Monotherapy against HIV has undoubted theoretical advantages and has good scientific fundaments. However, it is still controversial and here we will analyze the efficacy and safety of MT with darunavir with ritonavir (DRV/r) on patients who have received this treatment in our hospitals. MATERIALS AND METHODS Observational retrospective study that includes patients from 10 Andalusian hospitals that have received DRV/r in MT and that have been followed over a minimum of 12 months. We carried out a statistical descriptive analysis based on the profile of patients who had been prescribed MT and the efficacy and safety that were observed, paying special attention to treatment failure and virological evolution. RESULTS DRV/r was prescribed to 604 patients, of which 41.1% had a CD4 nadir <200/mmc. 33.1% had chronic hepatitis caused by HCV, had received an average of five lines of previous treatment and had a history of treatment failure to analogues in 33%, to non-analogues 22 and protease inhibitors (PI) in 19.5%. 76.6% proceeded from a previous treatment with PI. The simplification was the main criteria for the instauration of MT in the 81.5% and the adverse effects in the 18.5%. We managed to maintain MT in 84% of cases, with only 4.8% of virological failure (VF) with viral load (VL) >200 c/mL and 3.6% additional losses due to VF with VL between 50 and 200 copies/mL. Thirty three genotypes were performed after failure without findings of resistance mutations to DRV/r or other IPs. Only 23.7% of patients presented some blips during the period of exposition to MT. Eighty seven percent of all determinations of VL had <50 copies/mL, and only 4.99% had >200 copies/mL. Although up to 14.9% registered at some point an AE, only 2.6% abandoned MT because of AE and 1.2% because of voluntary decision. Although the average of total and LDL cholesterol increases 10 mg/dL after 2 years of follow-up, so did HDL cholesterol in 3mg/dL and the values of triglycerides (-14 mg/dL) and GPT (-6 UI/mL) decreased. The average count of CD4 lymphocytes increased from 642 to 714/mm(3) at 24 weeks. CONCLUSIONS In a very broad series of patients obtained from clinical practice, data from clinical trials was confirmed: MT with DRV as a de-escalation strategy is very safe, it's associated to a negligible rate of adverse effects and maintains a good suppression of HIV replication. VF (with >50 or >200 copies/mL) is always under 10% and in any case without consequences.

CNS safety at 48-week of switching to ATV/r plus 3TC or two nucleos(t)ides in HIV-suppressed patients on stable ART: the SALT neurocognitive sub-study.

INTRODUCTION Due to their low CNS penetrance, there are concerns about the capacity of non-conventional PI-based ART (monotherapy and dual therapies) to preserve neurocognitive performance (NP). METHODS We evaluated the NP change of aviremic participants of the SALT clinical trial (1) switching therapy to dual therapy (DT: ATV/r+3TC) or triple therapy (TT: ATV/r+2NRTI) who agreed to perform an NP assessment (NPZ-5) at baseline and W48. Neurocognitive impairment and NP were assessed using AAN-2007 criteria (2) and global deficit scores (GDS) (3). Neurocognitive change (GDS change: W48 - baseline) and the effect of DT on NP evolution crude and adjusted by significant confounders were determined using ANCOVA. RESULTS A total of 158 patients were included (Table 1). They had shorter times because HIV diagnosis, ART initiation and HIV-suppression and their virologic outcome at W48 by snapshot was higher (79.1% vs 72.7%; p=0.04) compared to the 128 patients not included in the sub-study. By AAN-2007 criteria, 51 patients in each ART group (68% vs 63%) were neurocognitively impaired at baseline (p=0.61). Forty-seven patients were not reassessed at W48: 30 lost of follow-up (16 DT-14 TT) and 17 had non-evaluable data (6 DT-11 TT). Patients retested were more likely to be men (78.9% vs 61.4%) and had neurological cofounders (9.6% vs 0%) than patients non-retested. At W48, 3 out of 16 (5.7%) patients on DT and 6 out of 21 (10.5%) on TT who were non-impaired at baseline became impaired (p=0.49) while 10 out of 37 (18.9%) on DT and 7 out of 36 (12.3%) on TT who were neurocognitively impaired at baseline became non-impaired (p=0.44). Mean GDS changes (95% CI) were: Overall -0.2 (-0.3 to -0.04): DT -0.26 (-0.4 to -0.07) and TT -0.08 (-0.2 to 0.07). NP was similar between DT and TT (0.15). This absence of differences was also observed in all cognitive tests. Effect of DT: -0.16 [-0.38 to 0.06]) (r(2)=0.16) on NP evolution was similar to TT (reference), even after adjusting (DT: -0.11 [-0.33 to 0.1], TT: reference) by significant confounders (geographical origin, previous ATV use and CD4 cell count) (r(2)=0.25). CONCLUSIONS NP stability was observed after 48 weeks of follow up in the majority of patients whether DT or TT was used to maintain HIV-suppression. Incidence rates of NP impairment or NP impairment recovery were also similar between DT and TT.

Should screening for Chlamydia trachomatis and Neisseria gonorrhoeae in HIV-men who have sex with men be recommended?

INTRODUCTION Sexually transmitted infections (STI) like Chlamydia trachomatis (CT) and Neisseria gonorrhoeae (NG) have been associated with increased risk of HIV acquisition (1). It has been also described as a high prevalence of asymptomatic CT and NG infections in men who have sex with men (MSM) (2). The aim of this study was to know the prevalence of CT and/or NG infections in asymptomatic HIV-MSM and the related factors. MATERIALS AND METHODS Prospective study of a cohort of asymptomatic HIV-MSM with follow-up in Malaga (southern Spain) during October 2012-May 2014. Patients with an opportunistic event or who received active antibiotic therapy for CT and/or NG in the previous month were excluded. All of them completed a questionnaire about sexual behaviour, barrier methods and recreational drugs use. Demographical, epidemiological, clinical, analytical and therapeutic data were also collected. Pharyngeal and rectal swabs, and urine samples were collected to be tested for CT and NG by nucleic acid amplification test (c4800 CT/NG. Roche Diagnostics, Mannheim, Germany) (3). STATISTICS ANALYSIS SPSS 17.0. RESULTS 255 patients were asked to participate and 248 of them accepted. Median age was 37.7 (30.6-46.3) years, median time since HIV diagnosis was 47.7 (10.5-104.1) months, and median CD4 cells count was 607 (440-824) cell/µL. There were 195 (78.6%) patients on antiretroviral therapy; 81.5% of them had undetectable viral load. 80.5% of the patients had a past history of STI. Infection by CT and/or NG was diagnosed in 24 (9.7%) patients. Overall four urine samples, two pharyngeal, and 15 rectal ones were positive for CT, and five pharyngeal and five rectal swabs were positive for NG. Two patients were co-infected by CT and NG: one with CT in urine and both in rectum, another with CT in urine and rectum and NG in pharynx. One patient presented CT in pharynx and rectum, and two patients NG in pharynx and rectum. Positive CT and/or NG tests were only related with detectable HIV viral load (OR 3.08, 95% CI 1.2-7.4; p=0.01). It was not related with sexual behaviour, nor with alcohol or recreational drugs use. CONCLUSIONS STI screening had a great acceptance in this population. There was a high prevalence of asymptomatic CT and/or NG infections. Rectum sample was the most effective one. Viral suppression could protect from these STI. Screening should be recommended in HIV-MSM.

Hepatitis E Virus seroprevalence and chronic infections in patients with HIV, Switzerland.

We screened 735 HIV-infected patients in Switzerland with unexplained alanine aminotransferase elevation for hepatitis E virus (HEV) immunoglobulin G. Although HEV seroprevalence in this population is low (2.6%), HEV RNA can persist in patients with low CD4 cell counts. Findings suggest chronic HEV infection should be considered as a cause of persistent alanine aminotransferase elevation.

HIV-1 coreceptor usage and CXCR4-specific viral load predict clinical disease progression during combination antiretroviral therapy

Background: Although combination antiretroviral therapy (cART) dramatically reduces rates of AIDS and death, a minority of patients experience clinical disease progression during treatment. <p>Objective: To investigate whether detection of CXCR4(X4)-specific strains or quantification of X4-specific HIV-1 load predict clinical outcome. Methods: From the Swiss HIV Cohort Study, 96 participants who initiated cART yet subsequently progressed to AIDS or death were compared with 84 contemporaneous, treated nonprogressors. A sensitive heteroduplex tracking assay was developed to quantify plasma X4 and CCR5 variants and resolve HIV-1 load into coreceptor-specific components. Measurements were analyzed as cofactors of progression in multivariable Cox models adjusted for concurrent CD4 cell count and total viral load, applying inverse probability weights to adjust for sampling bias. Results: Patients with X4 variants at baseline displayed reduced CD4 cell responses compared with those without X4 strains (40 versus 82 cells/mu l; P= 0.012). The adjusted multivariable hazard ratio (HR) for clinical progression was 4.8 [95% confidence interval (Cl) 2.3-10.0] for those demonstrating X4 strains at baseline. The X4-specific HIV-1 load was a similarly independent predictor, with HR values of 3.7(95%Cl, 1.2-11.3) and 5.9 (95% Cl, 2.2-15.0) for baseline loads of 2.2-4.3 and > 4.3 log(10)copies/ml, respectively, compared with < 2.2 log(10)copies/ml. Conclusions: HIV-1 coreceptor usage and X4-specific viral loads strongly predicted disease progression during cART, independent of and in addition to CD4 cell count or total viral load. Detection and quantification of X4 strains promise to be clinically useful biomarkers to guide patient management and study HIV-1 pathogenesis.

Clinicopathologic analysis of acute myeloid leukemia arising from chronic myelomonocytic leukemia.

Acute myeloid leukemia arising from chronic myelomonocytic leukemia is currently classified as acute myeloid leukemia with myelodysplasia-related changes, a high-risk subtype. However, the specific features of these cases have not been well described. We studied 38 patients with chronic myelomonocytic leukemia who progressed to acute myeloid leukemia. We compared the clinicopathologic and genetic features of these cases with 180 patients with de novo acute myeloid leukemia and 34 patients with acute myeloid leukemia following myelodysplastic syndromes. We also examined features associated with progression from chronic myelomonocytic leukemia to acute myeloid leukemia by comparing the progressed chronic myelomonocytic leukemia cases with a cohort of chronic myelomonocytic leukemia cases that did not transform to acute myeloid leukemia. Higher white blood cell count, marrow cellularity, karyotype risk score, and Revised International Prognostic Scoring System score were associated with more rapid progression from chronic myelomonocytic leukemia to acute myeloid leukemia. Patients with acute myeloid leukemia ex chronic myelomonocytic leukemia were older (P<0.01) and less likely to receive aggressive treatment (P=0.02) than de novo acute myeloid leukemia patients. Most cases showed monocytic differentiation and fell into the intermediate acute myeloid leukemia karyotype risk group; 55% had normal karyotype and 17% had NPM1 mutation. Median overall survival was 6 months, which was inferior to de novo acute myeloid leukemia (17 months, P=0.002) but similar to post myelodysplastic syndrome acute myeloid leukemia. On multivariate analysis of all acute myeloid leukemia patients, only age and karyotype were independent prognostic variables for overall survival. Our findings indicate that acute myeloid leukemia following chronic myelomonocytic leukemia displays aggressive behavior and support placement of these cases within the category of acute myeloid leukemia with myelodysplasia-related changes. The poor prognosis of these patients may be related to an older population and lack of favorable-prognosis karyotypes that characterize many de novo acute myeloid leukemia cases.

Bonafide, type-specific human papillomavirus persistence among HIV-positive pregnant women: predictive value for cytological abnormalities, a longitudinal cohort study

This study investigated the rate of human papillomavirus (HPV) persistence, associated risk factors, and predictors of cytological alteration outcomes in a cohort of human immunodeficiency virus-infected pregnant women over an 18-month period. HPV was typed through L1 gene sequencing in cervical smears collected during gestation and at 12 months after delivery. Outcomes were defined as nonpersistence (clearance of the HPV in the 2nd sample), re-infection (detection of different types of HPV in the 2 samples), and type-specific HPV persistence (the same HPV type found in both samples). An unfavourable cytological outcome was considered when the second exam showed progression to squamous intraepithelial lesion or high squamous intraepithelial lesion. Ninety patients were studied. HPV DNA persistence occurred in 50% of the cases composed of type-specific persistence (30%) or re-infection (20%). A low CD4+T-cell count at entry was a risk factor for type-specific, re-infection, or HPV DNA persistence. The odds ratio (OR) was almost three times higher in the type-specific group when compared with the re-infection group (OR = 2.8; 95% confidence interval: 0.43-22.79). Our findings show that bonafide (type-specific) HPV persistence is a stronger predictor for the development of cytological abnormalities, highlighting the need for HPV typing as opposed to HPV DNA testing in the clinical setting.

Study protocol for a pragmatic randomised controlled trial in general practice investigating the effectiveness of acupuncture against migraine.

BACKGROUND Migraine is a chronic neurologic disease that can severely affect the patient's quality of life. Although in recent years many randomised studies have been carried out to investigate the effectiveness of acupuncture as a treatment for migraine, it remains a controversial issue. Our aim is to determine whether acupuncture, applied under real conditions of clinical practice in the area of primary healthcare, is more effective than conventional treatment. METHODS/DESIGN The design consists of a pragmatic multi-centre, three-armed randomised controlled trial, complemented with an economic evaluation of the results achieved, comparing the effectiveness of verum acupuncture with sham acupuncture, and with a control group receiving normal care only. Patients eligible for inclusion will be those presenting in general practice with migraine and for whom their General Practitioner (GP) is considering referral for acupuncture. Sampling will be by consecutive selection, and by randomised allocation to the three branches of the study, in a centralised way following a 1:1:1 distribution (verum acupuncture; sham acupuncture; conventional treatment). Secondly, one patient in three will be randomly selected from each of the acupuncture (verum or sham) groups for a brain perfusion study (by single photon emission tomography). The treatment with verum acupuncture will consist of 8 treatment sessions, once a week, at points selected individually by the acupuncturist. The sham acupuncture group will receive 8 sessions, one per week, with treatment being applied at non-acupuncture points in the dorsal and lumbar regions, using the minimal puncture technique. The control group will be given conventional treatment, as will the other two groups. DISCUSSION This trial will contribute to available evidence on acupuncture for the treatment of migraine. The primary endpoint is the difference in the number of days with migraine among the three groups, between the baseline period (the 4 weeks prior to the start of treatment) and the period from weeks 9 to 12. As a secondary aspect, we shall record the index of laterality and the percentage of change in the mean count per pixel in each region of interest measured by the brain perfusion tomography, performed on a subsample of the patients within the real and sham acupuncture groups. TRIAL REGISTRATION Current Controlled Trials ISRCTN98703707.

Colistin resistance in a clinical Acinetobacter baumannii strain appearing after colistin treatment: effect on virulence and bacterial fitness.

The fitness and virulence costs associated with the clinical acquisition of colistin resistance by Acinetobacter baumannii were evaluated. The growth of strain CR17 (colistin resistant) was less than that of strain CS01 (colistin susceptible) when the strains were grown in competition (72-h competition index, 0.008). In a murine sepsis model, CS01 and CR17 reached spleen concentrations when coinfecting of 9.31 and 6.97 log10 CFU/g, respectively, with an in vivo competition index of 0.016. Moreover, CS01 was more virulent than CR17 with respect to mortality and time to death.

Quantifying rock fabrics: a test of autocorrelation of the spatial distribution of cristals

A novel test of spatial independence of the distribution of crystals or phases in rocksbased on compositional statistics is introduced. It improves and generalizes the commonjoins-count statistics known from map analysis in geographic information systems.Assigning phases independently to objects in RD is modelled by a single-trial multinomialrandom function Z(x), where the probabilities of phases add to one and areexplicitly modelled as compositions in the K-part simplex SK. Thus, apparent inconsistenciesof the tests based on the conventional joins{count statistics and their possiblycontradictory interpretations are avoided. In practical applications we assume that theprobabilities of phases do not depend on the location but are identical everywhere inthe domain of de nition. Thus, the model involves the sum of r independent identicalmultinomial distributed 1-trial random variables which is an r-trial multinomialdistributed random variable. The probabilities of the distribution of the r counts canbe considered as a composition in the Q-part simplex SQ. They span the so calledHardy-Weinberg manifold H that is proved to be a K-1-affine subspace of SQ. This isa generalisation of the well-known Hardy-Weinberg law of genetics. If the assignmentof phases accounts for some kind of spatial dependence, then the r-trial probabilitiesdo not remain on H. This suggests the use of the Aitchison distance between observedprobabilities to H to test dependence. Moreover, when there is a spatial uctuation ofthe multinomial probabilities, the observed r-trial probabilities move on H. This shiftcan be used as to check for these uctuations. A practical procedure and an algorithmto perform the test have been developed. Some cases applied to simulated and realdata are presented.Key words: Spatial distribution of crystals in rocks, spatial distribution of phases,joins-count statistics, multinomial distribution, Hardy-Weinberg law, Hardy-Weinbergmanifold, Aitchison geometry

Arqueología de la producción agrícola. Aplicación de técnicas de análisis estadístico y geocomputacional a un caso de estudio del IV-II milenio ANE en la Cuenca del Río Grande (Perú)

Muchas investigaciones arqueobotánicas, desde un enfoque cualitativo-descriptivo, limitan el propio campo de estudio a los análisis de presencia/ausencia y/o de frecuencia de taxones a partir de su recuento en los conjuntos vegetales. De esa manera, los datos proporcionados resultan ser inconcluyentes y no fiables para la reconstrucción del paleoambiente, la determinación de la dieta alimenticia y de la práctica económica realizada (recolección VS agricultura), y totalmente insuficientes para determinar los cambios históricos ocurridos en los procesos productivos. Por lo que concierne el Perú, desde los primeros estudios con referencia a restos vegetales recuperados en yacimientos arqueológicos, principalmente de la costa, se documenta el importante papel que han desarrollado las especies vegetales en la vida de las comunidades pre-hispánicas. No obstante la excepcional abundancia y óptima preservación de este tipo de material (botánico) en muchos de los yacimientos arqueológicos de esta región, gracias a las extremas condiciones climáticas y ambientales sobre todo de sus áridas zonas costeras, los estudios arqueobotánicos desarrollados hasta el momento son muy escasos y las limitaciones análiticas que presentan en su mayoría reflejan la poca importancia dada a las investigaciones arqueobotánicas. En el presente trabajo desarrollamos y aplicamos una metodología analítica de tipo cuantitativa para el estudio de los macrorestos vegetales procedentes de un yacimiento de la Costa sur del Perú. Con ello pretendemos obtener datos representativos y objetivos de los conjuntos analizados, cuyo procesado lleve a una exhaustiva y correcta interpretación de la información.

2

The pharmacokinetic profile of imatinib has been assessed in healthy subjects and in population studies among thousands of patients with CML or GIST. Imatinib is rapidly and extensively absorbed from the GI tract, reaching a peak plasma concentration (Cmax) within 1-4 h following administration. Imatinib bioavailability is high (98%) and independent of food intake. Imatinib undergoes rapid and extensive distribution into tissues, with minimal penetration into the central nervous system. In the circulation, it is approximately 95% bound to plasma proteins, principally α1-acid glycoprotein (AGP) and albumin. Imatinib undergoes metabolism in the liver via the cytochrome P450 enzyme system (CYP), with CYP3A4 being the main isoenzyme involved. The N-desmethyl metabolite CGP74588 is the major circulating active metabolite. The typical elimination half-life for imatinib is approximately 14-22 h. Imatinib is characterized by large inter-individual pharmacokinetic variability, which reflects in a wide spread of concentrations observed under standard dosage. Besides adherence, several factors have been shown to influence this variability, especially demographic characteristics (sex, age, body weight and disease diagnosis), blood count characteristics, enzyme activity (mainly CYP3A4), drug interactions, activity of efflux transporters and plasma levels of AGP. Additionally, recent retrospective studies have shown that drug exposure, reflected in either the area under the concentration-time curve (AUC) or more conveniently the trough level (Cmin), correlates with treatment outcomes. Increased toxicity has been associated with high plasma levels, and impaired clinical efficacy with low plasma levels. While no upper concentration limit has been formally established, a lower limit for imatinib Cmin of about 1000 ng/mL has been proposed repeatedly for improving outcomes in CML and GIST patients. Imatinib is licensed for use in chronic phase CML and GIST at a fixed dose of 400 mg once daily (600 mg in some other indications) despite substantial pharmacokinetic variability caused by both genetic and acquired factors. The dose can be modified on an individual basis in cases of insufficient response or substantial toxic effects. Imatinib would, however, meet traditional criteria for a therapeutic drug monitoring (TDM) program: long-term therapy, measurability, high inter-individual but restricted intra-individual variability, limited pharmacokinetic predictability, effect of drug interactions, consistent association between concentration and response, suggested therapeutic threshold, reversibility of effect and absence of early markers of efficacy and toxic effects. Large-scale, evidence-based assessments of drug concentration monitoring are therefore still warranted for the personalization of imatinib treatment.

Association of alcohol consumption and HIV surrogate markers in participants of the swiss HIV cohort study.

BACKGROUND: Alcohol consumption may affect the course of HIV infection and/or antiretroviral therapy (ART). The authors investigated the association between self-reported alcohol consumption and HIV surrogate markers in both treated and untreated individuals. DESIGN: Prospective cohort study. METHODS: Over a 7-year period, the authors analyzed 2 groups of individuals in the Swiss HIV Cohort Study: (1) ART-naïve individuals remaining off ART and (2) individuals initiating first ART. For individuals initiating first ART, time-dependent Cox proportional hazards models were used to assess the association between alcohol consumption, virological failure, and ART interruption. For both groups, trajectories of log-transformed CD4 cell counts were analyzed using linear mixed models with repeated measures. RESULTS: The authors included 2982 individuals initiating first ART and 2085 ART naives. In individuals initiating first ART, 241 (8%) experienced virological failure. Alcohol consumption was not associated with virological failure. ART interruption was noted in 449 (15%) individuals and was more prevalent in severe compared with none/light health risk drinkers [hazard ratio: 2.24, 95% confidence interval: 1.42 to 3.52]. The association remained significant even after adjusting for nonadherence. The authors did not find an association between alcohol consumption and change in CD4 cell count over time in either group. CONCLUSIONS: No effect of alcohol consumption on either virological failure or CD4 cell count in both groups of ART-initiating and ART-naive individuals was found. However, severe drinkers were more likely to interrupt ART. Efforts on ART continuation should be especially implemented in individuals reporting high alcohol consumption.