Machine learning is increasingly introduced into medical fields, yet there is limited evidence for its benefit over more commonly used statistical methods in epidemiological studies. We introduce an unsupervised machine learning framework for longitudinal features and evaluate it using sexual behaviour data from the last 20 years from over 3’700 participants in the Swiss HIV Cohort Study (SHCS). We use hierarchical clustering to find subgroups of men who have sex with men in the SHCS with similar sexual behaviour up to May 2017, and apply regression to test whether these clusters enhance predictions of sexual behaviour or sexually transmitted diseases (STIs) after May 2017 beyond what can be predicted with conventional parameters. We find that behavioural clusters enhance model performance according to likelihood ratio test, Akaike information criterion and area under the receiver operator characteristic curve for all outcomes studied, and according to Bayesian information criterion for five out of ten outcomes, with particularly good performance for predicting future sexual behaviour and recurrent STIs. We thus assess a methodology that can be used as an alternative means for creating exposure categories from longitudinal data in epidemiological models, and can contribute to the understanding of time-varying risk factors.

La mononucleose infectieuse aigue est un probleme frequent en medecine de ville. Dans 90% des cas, la mononucleose est due au virus Epstein-Barr (EBV) – pour de nombreux medecins, «EBV» et «mononucleose» sont synonymes. Le diagnostic n’est pas toujours facile, et certaines personnes atteintes ne presentent pas d’autres symptomes que la fatigue.

Die akute infektiose Mononukleose, oft auch Pfeiffersches Drusenfieber genannt, ist ein haufiges Problem in der Praxis. In 90% wird die Mononukleose durch das Epstein-Barr-Virus (EBV) verursacht – fur viele Arztinnen und Arzte sind «EBV» und «Pfeiffer» synonym. Die Diagnose ist nicht immer ganz einfach, zum Teil haben die Betroffenen ausser Mudigkeit wenig andere Symptome, und die Serologie ist nicht immer eindeutig.

Background. Coronary artery disease (CAD) events have been associated with certain antiretroviral therapy (ART) agents. In contrast, the influence of ART on subclinical atherosclerosis is not clear. The study objective was to assess the association between individual ART agents and the prevalence and extent of subclinical CAD. Methods. Coronary artery calcium (CAC) scoring and coronary computed tomography angiography (CCTA) were performed in >= 45-year-old Swiss Human Immunodeficiency Virus Cohort Study participants. The following subclinical CAD endpoints were analyzed separately: CAC score >0, any plaque, calcified plaque, noncalcified/mixed plaque, segment involvement score (SIS), and segment severity score (SSS). Logistic regression models calculated by inverse probability of treatment weights (IPTW) were used to explore associations between subclinical CAD and cumulative exposure to the 10 most frequently used drugs. Results. There were 403 patients who underwent CCTA. A CAC score >0 was recorded in 188 (47%), any plaque in 214 (53%), calcified plaque in 151 (38%), and noncalcified/mixed plaque in 150 (37%) participants. A CAC score >0 was negatively associated with efavirenz (IPTW adjusted odds ratio per 5 years 0.73, 95% confidence interval [CI] 0.56-0.96), tenofovir disoproxil fumarate (0.68, 95% CI 0.49-0.95), and lopinavir (0.64, 95% CI 0.43-0.96). Any plaque was negatively associated with tenofovir disoproxil fumarate (0.71, 95% CI 0.51-0.99). Calcified plaque was negatively associated with efavirenz (0.7, 95% CI 0.57-0.97). Noncalcified/mixed plaque was positively associated with abacavir (1.46, 95% CI 1.08-1.98) and negatively associated with emtricitabine (0.67, 95% CI 0.46-0.99). For SSS and SIS, we found no association with any drug. Conclusions. An increased risk of noncalcified/mixed plaque was only found in patients exposed to abacavir. Emtricitabine was negatively associated with noncalcified/mixed plaque, while tenofovir disoproxil fumarate and efavirenz were negatively associated with any plaque and calcified plaque, respectively.

Background: Whereas tenofovir disoproxil fumarate (TDF) can lead to renal adverse events, tenofovir alafenamide (TAF) has a more favorable renal safety profile. However, the impact of replacing TDF with TAF on renal function and liver parameters among HIV/hepatitis B virus (HBV)-coinfected individuals with renal dysfunction remains unclear. Methods: We included all participants from the Swiss HIV Cohort Study with an HIV/HBV coinfection who switched from TDF to TAF and had an estimated glomerular filtration rate (eGFR) <90 mL/min/1.73 m 2 and a suppressed HIV viral load (<200 cp/mL). We assessed changes in eGFR, urine protein-to-creatinine ratio, and alanine aminotransferase (ALT) after 1 year using mixed-effect models with interrupted time series. Results: Among 106 participants (15.1% women, median age 53 years), eGFR was 60–89 mL/min/1.73 m 2 in 84 (79.2%) and <60 mL/min/1.73 m 2 in 22 (20.8%) individuals at the time of switch. One year after the switch from TDF to TAF, individuals with an eGFR between 60 and 89 mL/min/1.73 m 2 experienced increases in eGFR of 3.2 mL/min/1.73 m 2 (95% confidence interval [CI] 1.2 to 5.2), whereas those with an eGFR <60 mL/min/1.73 m 2 experienced improvements of 6.2 mL/min/1.73 m 2 (95% CI 2.4 to 10.0). Urine protein-to-creatinine ratio decreased overall (−6.3 mg/mmol, 95% CI −10.0 to −2.7), and ALT levels declined in patients with elevated baseline levels (−11.8 IU/L, 95% CI −17.3 to −6.4) 1 year after replacing TDF with TAF. Conclusions: Switching from TDF to TAF among HIV/HBV-coinfected individuals with renal impairment led to improvements in eGFR, a decline in proteinuria, and to ALT normalization in those with elevated ALT levels.

Abstract Background This study tested a theory-based adherence-enhancing intervention: the “Interprofessional Medication Adherence Program” (IMAP) to increase human immunodeficiency virus (HIV) retention in care. Methods We retrospectively compared our intervention center (intervention group [IG]) with a standard of care center (control group [CG]) both participating in the Swiss HIV Cohort Study between 2004 and 2012. Endpoints were defined as >6-month and >12-month gaps in care for intervals of care longer than 6 and 12 months without any blood draw. Inverse probability of treatment weights was used to adjust for differences between patients at the 2 centers. Viral failure was defined as ribonucleic acid ≥50 copies/mL after 24+ weeks on antiretrovirals. Results The IG included 451 patients, CG 311. In the IG, 179 (40%) patients took part in the IMAP for a median of 27 months (interquartile range, 12–45). Gaps in care of ≥6 months were significantly more likely to happen in the CG versus IG (74.6% vs 57%, P < .001). The median time until the first treatment gap was longer in the IG vs CG (120 vs 84 weeks, P < .001). Gaps in care of ≥12 months evaluated in 709 (93%) patients were significantly more likely to occur in the CG compared with the IG (22.6% vs 12.5%, P < .001). The rate of viral failure was significantly lower in the IG (8.3% vs 15.1%, P = .003). Conclusions This study, in a real-world setting, shows the effectiveness of the IMAP to reduce 6- and 12-month gaps in follow up among people with HIV. These results should be confirmed by studies in other settings.

Background. Beta-hemolytic streptococci (BHS) are an uncommon cause of infective endocarditis (IE). The aim of this study was to describe the clinical features and outcomes of patients with BHS IE in a large multinational cohort and compare them with patients with viridans streptococcal IE. Methods. The International Collaboration on Endocarditis Prospective Cohort Study (ICE-PCS) is a large multinational database that recruited patients with IE prospectively using a standardized data set. Sixty-four sites in 28 countries reported patients prospectively using a standard case report form developed by ICE collaborators. Results. Among 1336 definite cases of streptococcal IE, 823 were caused by VGS and 147 by BHS. Patients with BHS IE had a lower prevalence of native valve (P < .005) and congenital heart disease predisposition (P = .002), but higher prevalence of implantable cardiac device predisposition (P < .005). Clinically, they were more likely to present acutely (P < .005) and with fever (P = .024). BHS IE was more likely to be complicated by stroke and other systemic emboli (P < .005). The overall in-hospital mortality of BHS IE was significantly higher than that of VGS IE (P = .001). In univariate analysis, variables associated with in-hospital mortality for BHS IE were age (odds ratio [OR], 1.044; P = .004), prosthetic valve IE (OR, 3.029; P = .022), congestive heart failure (OR, 2.513; P = .034), and stroke (OR, 3.198; P = .009). Conclusions. BHS IE is characterized by an acute presentation and higher rate of stroke, systemic emboli, and in-hospital mortality than VGS IE. Implantable cardiac devices as a predisposing factor were more often found in BHS IE compared with VGS IE.

Background. People with HIV (HIV+) may have increased cardiovascular event rates compared with HIV-negative (HIV-) persons. Cross-sectional data from the United States and Switzerland, based on coronary artery calcium scan (CAC) and coronary computed tomography angiography (CCTA), suggest, respectively, increased and similar prevalence of subclinical atherosclerosis in HIV+ vs HIV- persons. Methods. We repeated CAC/CCTA in 340 HIV+ and 90 HIV- study participants >2 years after baseline CAC/CCTA. We as- sessed the association of HIV infection, Framingham risk score (FRS), and HIV-related factors with the progression of subclinical atherosclerosis. Results. HIV+ were younger than HIV- participants (median age, 52 vs 56 years; P < .01) but had similar median 10-year FRS (8.9% vs 9.0%; P = .82); 94% had suppressed HIV viral load. In univariable and multivariable analyses, FRS was associated with the incidence rate ratio (IRR) of new subclinical atherosclerosis at the follow-up CAC/CCTA, but HIV infection was not: any plaque (adjusted IRR for HIV+ vs HIV- participants, 1.21; 95% CI, 0.62-2.35), calcified plaque (adjusted IRR for HIV+ vs HIV- participants, 1.06; 95% CI, 0.56-2), noncalcified/mixed plaque (adjusted IRR for HIV+ vs HIV- participants, 1.24; 95% CI, 0.69-2.21), and high-risk plaque (adjusted IRR for HIV+ vs HIV- participants, 1.46; 95% CI, 0.66-3.20). Progression of CAC score between baseline and follow-up CAC/CCTA was similar in HIV+ (median annualized change [interquartile range {IQR}], 0.41 [0-10.19]) and HIV- participants (median annualized change [IQR], 2.38 [0-16.29]; P = .11), as was progression of coronary segment severity score (HIV+: median annualized change [IQR], 0 [0-0.47]; HIV-: median annualized change [IQR], 0 [0-0.52]; P = .10) and coronary segment involvement score (HIV+: median annualized change [IQR], 0 [0-0.45]; HIV-: median annualized change [IQR], 0 [0-0.41]; P = 25). Conclusions. In this longitudinal CAC/CCTA study from Switzerland, Framingham risk score was associated with progression of subclinical atherosclerosis, but HIV infection was not.

BACKGROUND:Neurocognitive impairment (NCI) in people with human immunodeficiency virus (PWH) remains a concern despite potent antiretroviral therapy (ART). Higher central nervous system (CNS) penetration effectiveness (CPE) scores have been associated with better CNS human immunodeficiency virus (HIV) replication control, but the association between CPE and NCI remains controversial. METHODS:The Neurocognitive Assessment in the Metabolic and Aging Cohort (NAMACO) study is a subgroup of the Swiss HIV Cohort Study (SHCS) that invited patients aged ≥45 years enrolled in the SHCS and followed-up at NAMACO-affiliated centers in Switzerland to participate between May 2013 and November 2016. In total, 981 patients were enrolled, all of whom underwent standardized neurocognitive assessment. Neurocognitive impairment, if present, was characterized using Frascati criteria. The CPE scores of NAMACO study participants with undetectable plasma HIV-ribonucleic acid at enrollment (909 patients) were analyzed. Cross-sectional CPE scores (at neurocognitive assessment) were examined as potential predictors of NCI in multivariate logistic regression models. The analysis was then repeated taking CPE as a cumulative score (summarizing CPE scores from ART initiation to the time of neurocognitive assessment). RESULTS:Most patients were male (80%) and Caucasian (92%). Neurocognitive impairment was present in 40%: 27% with HIV-associated NCI (mostly asymptomatic neurocognitive impairment), and 13% with NCI related to other factors. None of the CPE scores, neither cross-sectional nor cumulative, was statistically significantly associated with NCI. CONCLUSIONS:In this large cohort of aviremic PWH, we observed no association between NCI, whether HIV-associated or related to other factors, and CPE score, whether cross-sectional or cumulative.

Background. Self-reported neurocognitive impairment (SRNI) in people living with human immunodeficiency virus type 1 (HIV-1) infection is frequent. We use longitudinal information on SRNI in the Swiss HIV Cohort Study (SHCS) to identify and characterize groups of patients with persisting SRNI over time. Methods. We included all SHCS patients who were assessed for SRNI during at least 5 visits spanning at least 2.5 years in 2013-2017. We first compared patients with SRNI to those without SRNI over the whole study period. Second, we used a hierarchical cluster algorithm to identify groups of patients with similar changes of SRNI over time. In both analyses, we studied clinical and demographic factors potentially influencing SRNI. Results. In total, 79 683 questionnaires of 11 029 patients contained information about SRNI, and 8545 of 11 029 (77.5%) patients had longitudinal information. The overall percentage of patients with SRNI decreased from 19.6% in 2013 to 10.7% in 2017. Compared to patients in the cluster with low-level SRNI over time, patients in the cluster with high-level persisting SRNI more often had a prior opportunistic infection of the central nervous system (CNS) (odds ratio [OR], 3.7; P < .001), imperfect adherence to antiretroviral therapy (ART) (OR, 2.8; P < .001), and depression (OR, 1.9; P < .001). Conclusions. Although overall SRNI is decreasing in the SHCS, there is a group of patients with persisting SRNI over time. Past opportunistic infections of the CNS, imperfect adherence to ART, and depression were associated most with persisting SRNI. Patients with these characteristics should be preferentially tested for neurocognitive impairment.

Aims HIV-positive persons have increased cardiovascular event rates but data on the prevalence of subclinical atherosclerosis compared with HIV-negative persons are not uniform. We assessed subclinical atherosclerosis utilizing coronary artery calcium (CAC) scoring and coronary computed tomography angiography (CCTA) in 428 HIV-positive participants of the Swiss HIV Cohort Study and 276 HIV-negative controls concurrently referred for clinically indicated CCTA. Methods and results We assessed the association of HIV infection, cardiovascular risk profile, and HIV-related factors with subclinical atherosclerosis in univariable and multivariable analyses. HIV-positive participants (median duration of HIV infection, 15 years) were younger than HIV-negative participants (median age 52 vs. 56 years; P < 0.01) but had similar median 10-year Framingham risk scores (9.0% vs. 9.7%; P = 0.40). The prevalence of CAC score >0 (53% vs. 56.2%; P = 0.42) and median CAC scores (47 vs. 47; P = 0.80) were similar, as was the prevalence of any, non-calcified/mixed, and high-risk plaque. In multivariable adjusted analysis, HIV-positive participants had a lower prevalence of calcified plaque than HIV-negative participants [36.9% vs. 48.6%, P < 0.01; adjusted odds ratio (aOR) 0.57; 95% confidence interval (CI) 0.40-0.82; P < 0.01], lower coronary segment severity score (aOR 0.72; 95% CI 0.53-0.99; P = 0.04), and lower segment involvement score (aOR 0.71, 95% CI 0.52-0.97; P = 0.03). Advanced immunosuppression was associated with non-calcified/mixed plaque (aOR 1.97; 95% CI 1.09-3.56; P = 0.02). Conclusion HIV-positive persons in Switzerland had a similar degree of non-calcified/mixed plaque and high-risk plaque, and may have less calcified coronary plaque, and lower coronary atherosclerosis involvement and severity scores than HIV-negative persons with similar Framingham risk scores.

ObjectivesDolutegravir (DTG) is a highly effective integrase inhibitor with a strong genetic resistance barrier and a potential role in simplified HIV maintenance treatment. We assessed the feasibility of DTG maintenance monotherapy and measured HIV reservoirs on DTG monotherapy.MethodsAn interventional, open‐label, single‐arm study including eight virologically suppressed HIV‐1‐infected patients switched to DTG 50 mg once daily for 24 weeks was performed. HIV‐1 RNA levels in plasma and cerebrospinal and seminal fluids were measured at baseline and week 24, as well as HIV‐1 DNA in peripheral cells and DTG concentrations in these compartments.ResultsHIV‐1 RNA remained undetectable in all samples of blood, cerebrospinal fluid and sperm throughout the 24 weeks, except for one cerebrospinal fluid sample with a value of 28 HIV‐1 RNA copies/mL at week 24. One patient discontinued the study because of a neurological side effect. There was no change in the mean HIV‐1 DNA level between baseline and week 24. Plasma and cerebrospinal fluid DTG concentrations reached therapeutic levels in all patients in these two compartments.ConclusionsIn this small sample of carefully selected patients, HIV‐1 reservoirs were well controlled on DTG monotherapy over a period of 24 weeks. Viral suppression was also maintained throughout follow‐up.

Objectives: Time to blood culture positivity (TTP), a routinely available parameter in automated blood culture systems, may be a proxy for infectious burden in patients with bloodstream infections. We aimed to study the association between TTP and infective endocarditis (IE), or death, in patients with Staphylococcus aureus bacteraemia. Methods: VIRSTA is a multicentre prospective cohort study that included all adult patients with S. aureus bacteraemia in eight university hospitals in France (2009-2011). We analysed data from four centres which collected data on TTP. Regression models were used to study the association between TTP and definite IE (Duke-Li criteria), and 30 day-mortality. Results: We included 587 patients with S. aureus bacteraemia: mean age was 65.3 +/- 16.3 years, 420 out of 587 patients (71.6%) were male, 121 out of 587 (20.6%) died, and 42 out of 587 (7.2%) had definite IE. Median TTP of first positive blood culture was 13.7 h (interquartile range 9.9-18). On multivariate analysis, 30-day mortality was associated with TTP <= 13.7 h (74/295 (25.1%) vs. 47/292 (16.1%), p 0.02), as well as old age, McCabe score, methicillin resistance, stroke, pneumonia, and C-reactive protein. TTP was also independently associated with IE, but with a U-shape curve: IE was more common in the first (TTP <10 h, 17/148, 11.5%), and the last (TTP >= 18 h, 8/146, 5.5%) quartiles of TTP, p 0.002. Conclusions: TTP provides reliable information in patients with S. aureus bacteraemia, on the risk of IE, and prognosis, with short TTP being an independent predictor of death. These data, readily available at no cost, may be used to identify patients who require specific attention. (C) 2018 European Society of Clinical Microbiology and Infectious Diseases. Published by Elsevier Ltd. All rights reserved.

Abstract Background Hypercholesterolemia is a well established risk factor for coronary heart disease and is highly prevalent among human immunodeficiency virus (HIV)-positive persons. Antiretroviral therapy (ART) can both directly modify total cholesterol and have drug-drug interactions with statins. This makes investigating modifiable behavioral predictors of total cholesterol a pertinent task. Methods To explore the association between diet and physical activity with cross-sectionally measured total cholesterol, we administered a validated Food-Frequency-Questionnaire to participants of the Swiss HIV Cohort Study ≥45 years old. Linear mixed-effects models were applied to explore the associations between dietary patterns and physical activity with total cholesterol, after adjustment for clinical and demographic covariates. Results In total, 395 patients were included. Forty percent (158 of 395) had elevated total cholesterol (>5.2 mmol/L), and 41% (164 of 395) were not regularly physically active. In multivariable analysis, 2 factors were positively associated with total cholesterol; female sex (β = 0.562; 95% confidence interval [CI], 0.229–0.896) and the combined consumption of meat, refined/milled grains, carbonated beverages, and coffee (β = 0.243; 95% CI, 0.047–0.439). On the other hand, regular physical activity (β = −0.381; 95% CI, −0.626 to −0.136), lipid-lowering drugs (β = −0.443; 95% CI −0.691 to −0.196), ART containing tenofovir (β = −0.336; 95% CI −0.554 to −0.118), and black ethnicity (β = −0.967; 95% CI −1.524 to −0.410) exhibited a negative association. Conclusions We found independent associations between certain dietary patterns and physical activity with total cholesterol. Increasing physical activity might achieve cardiovascular and other health benefits in HIV-positive individuals. The clinical relevance of the identified dietary patterns requires further investigation in prospective cohort studies and randomized controlled trials.

Objective A ntiphospholipid (aPL) antibodies may activate platelets and contribute to vegetation growth and embolisation in infective endocarditis (IE). We aimed to determine the value of aPL as predictors of embolic events (EE) in IE. Methods We studied 186 patients with definite IE (Duke-Li criteria, all types of IE) from the Nanc-IE prospective registry (2007-2012) who all had a frozen blood sample and at least one imaging procedure to detect asymptomatic or confirm symptomatic EE. Anticardiolipin (aCL) and anti-beta(2) -glycoprotein I (beta(2) GPI) antibodies (IgG and IgM) were assessed after the end of patients' inclusion. The relationship between antibodies and the detection of EE after IE diagnosis were studied with Kaplan-Meier and Cox multivariate analyses. Results At least one EE was detected in 118 (63%) patients (52 cerebral, 95 other locations) after IE diagnosis in 80 (time interval between IE and EE diagnosis: 5.9 +/- 11.3 days). At least one aPL antibody was found in 31 patients (17%). Detection of EE over time after IE diagnosis was more frequent among patients with anti-beta(2) GPI IgM (log-rank P=0.0036) and that of cerebral embolisms, among patients with aCL IgM and anti-beta(2) GPI IgM (log-rank P=0.002 and P<0.0001, respectively). Factors predictive of EE were anti-beta(2) GPI IgM (HR=3.45 (1.47-8.08), P=0.0045), creatinine (2.74 (1.55-4.84), P=0.0005)and vegetation size (2.41 (1.41-4.12), P=0.0014). Those of cerebral embolism were aCL IgM (2.84 (1.22-6.62), P=0.016) and anti-beta(2) GPI IgM (4.77 (1.79-12.74), P=0.0018). Conclusion The presence of aCL and anti-beta(2) GPI IgM was associated with EE, particularly cerebral ones, and could contribute to assess the embolic risk of IE.

Le dépistage de l’infection par le VIH fait partie de la politique de prévention de l’infection par le VIH. Elle repose sur la responsabilisation individuelle et l’incitation au dépistage volontaire. En France des centres de dépistages anonymes et gratuits ont été mis en place pour faciliter cette démarche. Nous rapportons ici l’expérience d’un centre de dépistage anonyme mais non gratuit de l’infection par le VIH. L’inscription et la prise de rendez-vous pour le dépistage se font en ligne sur un site dédié, en même temps qu’un questionnaire anonyme. Le test est effectué à l’hôpital par une infirmière qui l’accompagne par du counseling. Celle-ci vérifie et complète si besoin avec le patient, le questionnaire. Avant 2006, les résultats étaient rendus la semaine suivante, lors d’un 2e rendez-vous. Après 2006, ils sont rendus immédiatement (méthode Determine). Nous présentons ici les questionnaires collectés entre le 1er janvier 2012 et le 31 décembre 2016. Parmi 23071 inscriptions en ligne, 17 817 ont été honorées : hommes 61 %, âge moyen : 32 ans (extrêmes 14–83 ans). La situation maritale des personnes était : célibat (60 %), en couple (18 %), séparés (12 %). Le motif du test était : probable exposition au VIH (44 %), nouvelle relation (35 %), peur (47 %), rupture de préservatif (12,5 %), abandon du préservatif (19 %), autres raisons (9 %). C’était le 1er test pour 25 % des personnes, le 2e ou plus pour 63 %. Les pratiques sexuelles étaient : relations avec des femmes (53 %), des hommes (39 %), les deux à la fois (4,5 %). Le nombre de partenaires durant la dernière année était > 1 pour 68,5 % des personnes (> 5 pour 12 %). Le port de préservatif était constant pour 53,5 %, absent ou rare pour 31 % avec un partenaire occasionnel. Par rapport à la population résidante, on peut estimer qu’environ 8 % (3556/44 865 personnes) de la population (âge moyen entre 30 et 35 ans) se sont faits tester en 2015. La consultation payante ne dissuade pas la population sexuellement active de se faire dépister. Celle-ci est majoritairement célibataire, se dépiste en raison d’une probable prise de risque et de la peur d’avoir été contaminé, pratique des tests régulièrement, et a de multiples partenaires principalement féminines. La majorité des personnes déclare utiliser toujours un préservatif lors des relations avec partenaires occasionnels. L’inscription anonyme en ligne incite probablement au dépistage.

Purpose: The implementation of genomic-based medicine is hindered by unresolved questions regarding data privacy and delivery of interpreted results to health-care practitioners. We used DNA-based prediction of HIV-related outcomes as a model to explore critical issues in clinical genomics.Methods: We genotyped 4,149 markers in HIV-positive individuals. Variants allowed for prediction of 17 traits relevant to HIV-medical care, inference of patient ancestry, and imputation of human leukocyte antigen (HLA) types. Genetic data were processed under a privacy-preserving framework using homomorphic encryption, and clinical reports describing potentially actionable results were delivered to health-care providers.Results: A total of 230 patients were included in the study. We demonstrated the feasibility of encrypting a large number of genetic markers, inferring patient ancestry, computing monogenic and polygenic trait risks, and reporting results under privacy-preserving conditions. The average execution time of a multimarker test on encrypted data was 865 ms on a standard computer. The proportion of tests returning potentially actionable genetic results ranged from 0 to 54%.Conclusions: The model of implementation presented herein informs on strategies to deliver genomic test results for clinical care. Data encryption to ensure privacy helps to build patient trust, a key requirement on the road to genomic-based medicine.

Aims Indications for surgery in acute infective endocarditis (IE) are detailed in guidelines, but their application is not well known. We analysed the agreement between the patient's attending physicians and European Society of Cardiology guidelines regarding indications for surgery. We also assessed whether surgery was performed in patients who had an indication.Methods and results From the 2008 prospective population-based French survey on IE, 303 patients with definite left-sided native IE were identified. For each case, we prospectively recorded (i) indication for surgery according to the attending physicians and (ii) indication for surgery according to guidelines. Surgery was indicated in 194 (65%) patients according to attending physicians and in 221 (73%) according to guidelines, while 139 (46%) underwent surgery. Agreement was moderate between attending physicians and guidelines (kappa 0.41-0.59) and between indication according to guidelines and the performance of surgery (kappa 0.38). Of the 90 (30%) patients not operated despite indication, contraindication to surgery was reported by the attending physicians in 42 (47%), and indication was not identified in 48 (53%). One-year survival was 76% in patients with indication and surgery performed (n = 131), 69% in patients without indication and no surgery (n = 74), 56% in patients with identified indication and contraindication to surgery (n = 42), and 60% in patients with no identified indication (n = 48; P = 0.059).Conclusion Cardiac surgery during acute IE was recommended in almost three out of four patients, although fewer than half were actually operated. Indication was not acknowledged by the attending physicians in one out of six patients.