Calcul de taille d’étude pour le développement d’un modèle de prédiction d’une aggravation d’une forme légère de COVID-19

Introduction

Le calcul de la taille d’étude pour le développement d’un modèle de prédiction clinique utilise encore souvent la règle empirique proposée par Peduzzi en 1996 qui recommande l’inclusion de 10 événements à prédire par paramètre envisagé dans le modèle de prédiction. Une nouvelle approche a été proposée par Riley et al. en 2020 pour répondre aux enjeux majeurs du développement d’un modèle de prédiction : précision des estimations et contrôle de l’optimisme sur les performances de prédiction. Nous rapportons ici l’application de cette approche pour le développement d’un modèle de prédiction combinant des facteurs cliniques, biologiques et d’imagerie pour prédire la survenue d’une aggravation d’une forme légère de COVID-19 chez des patients ayant des résultats d’imagerie thoracique compatibles ou typiques de la COVID-19.

Méthodes

La stratégie de modélisation envisagée était un modèle de régression logistique avec une sélection a priori de facteurs identifiés dans la littérature. Comme dans la règle empirique, la nouvelle approche nécessite de préciser l’incidence de l’événement à prédire et le nombre de paramètres envisagés dans le modèle de prédiction. La taille d’étude déterminée par cette approche doit satisfaire quatre critères : précision souhaitée pour l’incidence de l’événement ; précision souhaitée pour les prédictions individuelles (sous la forme d’un pourcentage d’erreur absolu moyen) ; niveau maximal souhaité de sur-ajustement aux données (facteur de réduction minimal) et d’optimisme (surestimation maximale du R² de Nagelkerke). Pour les deux derniers critères, le calcul requiert de préciser le R² de Cox–Snell qui, en l’absence d’éléments probants dans la littérature, peut être déterminé en faisant une hypothèse sur le niveau de performance attendu du modèle de prédiction en termes de pourcentage de la variation totale observée dans les données expliqué par le modèle.

Résultats

L’incidence de l’évènement à prédire a été fixée à 20 % selon les données de la littérature. Une revue de la littérature sur les modèles de prédiction développés dans le contexte d’infection à COVID-19 nous a conduits à envisager 16 paramètres dans le modèle. Nous avons suivi les recommandations des auteurs pour définir la précision souhaitée pour l’incidence d’événement (±0,05), le pourcentage d’erreur absolu moyen (5 %), le facteur de réduction minimal (³ 0,9) et la surestimation maximale du critère de performance (£ 0,05). Nous avons fait varier le pourcentage de la variation totale observée dans les données expliqué par le modèle entre 10 et 50 %. Pour satisfaire l’ensemble des critères, la taille d’étude requise variait entre 587 et 2205 patients. Nous avons retenu l’hypothèse d’un pourcentage de variance expliqué de 25 % conduisant à l’inclusion de 829 patients pour le développement du modèle.

Conclusion

La nouvelle approche de calcul de taille d’étude permet de considérer dès la conception, des éléments majeurs qui favorisent le développement d’un modèle de prédiction clinique de qualité, en particulier la précision des probabilités individuelles de survenue de l’événement et le contrôle de l’optimisme. En contrepartie, elle nécessite une expertise statistique plus poussée que pour la règle empirique. Néanmoins, son application est possible même dans un délai court et devrait donc être systématiquement envisagée dans les études de développement d’un modèle de prédiction clinique.

Risk factors for colonization and infection by Pseudomonas aeruginosa in patients hospitalized in intensive care units in France

Pseudomonas aeruginosa (P.aeruginosa) remains a prominent nosocomial pathogen responsible for high morbi-mortality in intensive care units (ICUs). P.aeruginosa transmission is known to be partly endogenous and exogenous. Main factors have been highlighted but the precise role of environment in regard to antibiotics use remained unclear.

Public Health and Epidemiology Informatics

Objectives: To summarize current research in the field of Public Health and Epidemiology Informatics. Methods: The complete 2016 literature concerning public health and epidemiology informatics has been searched in PubMed and Web of Science, and the returned references were reviewed by the two section editors to select 14 candidate best papers. These papers were then peer-reviewed by external reviewers to allow the editorial team an enlightened selection of the best papers. Results: Among the 829 references retrieved from PubMed and Web of Science, three were finally selected as best papers. The first one compares Google, Twitter, and Wikipedia as tools for Influenza surveillance. The second paper presents a Geographic Knowledge-Based Model for mapping suitable areas for Rift Valley fever transmission in Eastern Africa. The last paper evaluates the factors associated with the visit of Facebook pages devoted to Public Health Communication. Conclusions: Surveillance is still a productive topic in public health informatics but other very important topics in public health are appearing.

Reference curves for CD4 T-cell count response to combination antiretroviral therapy in HIV-1-infected treatment-naïve patients

OBJECTIVES

The aim of this work was to provide a reference for the CD4 T-cell count response in the early months after the initiation of combination antiretroviral therapy (cART) in HIV-1-infected patients.

METHODS

All patients in the Collaboration of Observational HIV Epidemiological Research Europe (COHERE) cohort who were aged ≥ 18 years and started cART for the first time between 1 January 2005 and 1 January 2010 and who had at least one available measurement of CD4 count and a viral load ≤ 50 HIV-1 RNA copies/mL at 6 months (± 3 months) after cART initiation were included in the study. Unadjusted and adjusted references curves and predictions were obtained using quantile regressions.

RESULTS

A total of 28 992 patients were included in the study. The median CD4 T-cell count at treatment initiation was 249 [interquartile range (IQR) 150, 336] cells/μL. The median observed CD4 counts at 6, 9 and 12 months were 382 (IQR 256, 515), 402 (IQR 274, 543) and 420 (IQR 293, 565) cells/μL. The two main factors explaining the variation of CD4 count at 6 months were AIDS stage and CD4 count at cART initiation. A CD4 count increase of ≥ 100 cells/mL is generally required in order that patients stay 'on track' (i.e. with a CD4 count at the same percentile as when they started), with slightly higher gains required for those starting with CD4 counts in the higher percentiles. Individual predictions adjusted for factors influencing CD4 count were more precise.

CONCLUSIONS

Reference curves aid the evaluation of the immune response early after antiretroviral therapy initiation that leads to viral control.

Early initiation of combined antiretroviral therapy preserves immune function in the gut of HIV-infected patients

Massive loss of lamina propria CD4(+) T cells, changes in the lymphatic architecture, and altered intestinal epithelial barrier leading to microbial translocation are the common features of HIV-1 infection and are not fully restored under combined antiretroviral therapy (cART). To better understand determinants of gut mucosal restoration, we have performed phenotypic and gene expression analyses of the gut from HIV-infected patients, naive or treated with cART initiated either at the early phase of the primary infection or later during the chronic phase. We found a depletion of T helper type 22 (Th22) and interleukin-17-producing cells in naive patients. These populations, except Th22 cells, were not restored under cART. Regulatory T cells/Th17 ratio was significantly increased in HIV-infected patients and was inversely correlated to the restoration of CD4(+) T cells but not to gut HIV DNA levels. Gene profile analysis of gut mucosal distinguished two groups of patients, which fitted with the timing of cART initiation. In their majority early, but not later treated patients, exhibited conserved intestinal lymphoid structure, epithelial barrier integrity and dendritic cell maturation pathways. Our data demonstrate that early initiation of cART helps to preserve and/or restore lymphoid gut mucosal homeostasis and provide a rationale for initiating cART during the acute phase of HIV infection.

A new approach to integrate toxicity grade and repeated treatment cycles in the analysis and reporting of phase I dose-finding trials

BACKGROUND

Safety assessment beyond the dose-limiting toxicity evaluation period provides relevant information to define the recommended phase II dose (RP2D) of a new treatment. We retrospectively analyzed three phase I trials to illustrate two indicators: per-cycle probability of graded toxicity and cumulative probability of severe toxicity over the treatment period.

PATIENTS AND METHODS

Data were collected from two continual reassessment method (CRM) trials (T1: aviscumine in solid tumors with short time on treatment; T2: erlotinib + radiotherapy in brainstem gliomas with longer time on treatment) and one 3 + 3 design (T3: liposomal doxorubicin + cyclophosphamide combination in ovarian carcinoma). The probability of severe and moderate or severe toxicity per cycle was estimated at each dose level with mixed proportional odds model. The cumulative probability of severe toxicity was also estimated with the time-to-event CRM.

RESULTS

Eighty-three patients were included in the three trials; 94, 96 and 72 treatment cycles were administered, in T1, T2 and T3, respectively. Moderate toxicities were at least twice as frequent as severe toxicities. An increased probability of toxicity over time was detected in T3 [P = 0.04; per-cycle probability of severe toxicity: 27% (cycle 1) to 59% (cycle 6) at the RP2D]. At the RP2D, 37% of patients experienced at least one severe toxicity over the first six cycles in T2, and 78% in T3.

CONCLUSIONS

Dedicated methods can be used to analyze toxicities from all cycles of treatment. They do not delay accrual and should be integrated in the analysis and reporting of phase I dose-finding trials.

Risk factors for Pseudomonas aeruginosa acquisition in intensive care units: a prospective multicentre study

BACKGROUND

Pseudomonas aeruginosa is a major nosocomial pathogen in intensive care units (ICUs); however, endogenous versus exogenous origin of contamination remains unclear.

AIM

To identify individual and environmental ICU risk factors for P. aeruginosa acquisition.

METHODS

A five-month prospective multicentric study was performed in ten French ICUs. Adult patients hospitalized in ICU for ≥ 24 h were included and screened for P. aeruginosa colonization on admission, weekly and before discharge. P. aeruginosa acquisition was defined by a subsequent colonization or infection if screening swabs on admission were negative. Water samples were obtained weekly on water taps of the ICUs. Data on patient characteristics, invasive devices exposure, antimicrobial therapy, P. aeruginosa water and patient colonization pressures, and ICU characteristics were collected. Hazard ratios (HRs) were estimated using multivariate Cox model.

FINDINGS

Among the 1314 patients without P. aeruginosa on admission, 201 (15%) acquired P. aeruginosa during their ICU stay. Individual characteristics significantly associated with P. aeruginosa acquisition were history of previous P. aeruginosa infection or colonization, cumulative duration of mechanical ventilation and cumulative days of antibiotics not active against P. aeruginosa. Environmental risk factors for P. aeruginosa acquisition were cumulative daily ward 'nine equivalents of nursing manpower use score' (NEMS) [hazard ratio (HR): 1.47 for ≥ 30 points; 95% confidence interval (CI): 1.06-2.03] and contaminated tap water in patient's room (HR: 1.76; CI: 1.09-2.84).

CONCLUSION

Individual risk factors and environmental factors for which intervention is possible were identified for P. aeruginosa acquisition.

Dynamical models of biomarkers and clinical progression for personalized medicine: the HIV context

Mechanistic models, based on ordinary differential equation systems, can exhibit very good predictive abilities that will be useful to build treatment monitoring strategies. In this review, we present the potential and the limitations of such models for guiding treatment (monitoring and optimizing) in HIV-infected patients. In the context of antiretroviral therapy, several biological processes should be considered in addition to the interaction between viruses and the host immune system: the mechanisms of action of the drugs, their pharmacokinetics and pharmacodynamics, as well as the viral and host characteristics. Another important aspect to take into account is clinical progression, although its implementation in such modelling approaches is not easy. Finally, the control theory and the use of intrinsic properties of mechanistic models make them very relevant for dynamic treatment adaptation. Their implementation would nevertheless require their evaluation through clinical trials.

Prevalence and factors associated with renal impairment in HIV-infected patients, ANRS C03 Aquitaine Cohort, France

OBJECTIVES

The aims of the present study were to estimate the prevalence of renal impairment (RI) among HIV-infected adult patients and to investigate the associated factors.

METHODS

A cross-sectional survey was conducted in a French hospital-based cohort. Clearance of creatinine (CC) was calculated using the Cockcroft-Gault formula. Four stages of RI were defined: mild (60-90 mL/min), moderate (30-60), severe (15-30) and end stage (<15). Logistic regression models were used to investigate factors associated with RI.

RESULTS

The male/female ratio of the 2588 patients enrolled was 3:1 and the median age was 42 years. At the time of assessment of CC, the median CD4 count was 430 cells/microL and HIV plasma viral load (VL) was<50 copies/mL in 60%. The overall prevalence of RI was 39.0%: 34.2% mild, 4.4% moderate, 0.3% severe and 0.2% end-stage. Mild RI was associated with female gender [odds ratio (OR)=3.3: 95% CI 2.6-4.3)], age >50 years (OR=9.8: 7.4-13.0) and 40-50 years (OR=1.9: 1.5-2.4), body mass index (BMI) <22 kg/m(2) (OR=3.3: 2.7-4.3) and tenofovir exposure (OR=1.4: 1.0-1.9 for <1 year and OR=1.5: 1.2-2.0 for >1 year). Advanced RI (CC <60 mL/min) was associated with age >50 years (OR=5.6: 2.9-10.9) and 40-50 years (OR=2.2: 1.1-1.4), BMI <22 kg/m(2) (OR=1.5: 1.0-2.4), hypertension (OR=2.5: 1.4-2.5) and indinavir (IDV) exposure >1 year (OR=2.3: 1.5-3.6).

CONCLUSION

This survey confirms the high prevalence of RI in HIV-infected patients and indicates the importance of the investigation of renal function especially in women, older patients, those with a low BMI or treated with tenofovir or IDV.

TNFSF15 polymorphisms are associated with susceptibility to inflammatory bowel disease in a new European cohort

OBJECTIVES

Inflammatory bowel disease (IBD), e.g., Crohn's disease (CD) and ulcerative colitis (UC), is a complex genetic disorder. Tumor necrosis factor (ligand) superfamily, member 15 (TNFSF15) has been previously identified as a susceptibility gene for CD in Japanese and UK cohorts. This replication study was designed in order to confirm and further validate the role of TNFSF15 in IBD.

METHODS

A total of 666 IBD families (corresponding to 2,982 relatives) with European ancestry were genotyped for the rs6478108 and rs7869487 polymorphisms, which define the main TNFSF15 haplotypes previously associated with CD. An association between the main haplotypes and CD, UC and IBD was tested using the Genehunter TDT and Unphased statistics. Caspase recruitment domain 15 (CARD15)/TNFSF15 interaction and genotype/phenotype correlations were also studied.

RESULTS

The previously reported "high-risk" haplotype (A) was associated with IBD (P=0.001) (OR=1.25 (1.05-1.50)) and CD (P=0.02) (OR=1.31 (1.03-1.67)) whereas the "protective" (B) haplotype was significantly less transmitted to IBD and CD patients. No interaction between CARD15 and TNFSF15 was detected. We also failed to define a clinical subgroup of CD patients specifically associated with TNFSF15 haplotype A.

CONCLUSIONS

This study confirms that TNFSF15 or a closely linked gene is involved in the genetic predisposition to CD.

When it is better to estimate a slope with only one point

When investigating the change in a biomarker, it is often believed that at least two measurements are needed from each participant, and that those with only one measurement should be excluded. In this short note, we explain why this could lead to imprecise and biased estimates. Furthermore, we discuss a standard statistical method that handles such issues.

CD4+ T-lymphocytes natural decrease in HAART-naïve HIV-infected adults in Abidjan

OBJECTIVE

To study the CD4 natural decrease and its determinants in sub-Saharan African HIV-infected adults.

METHOD

We performed a 7-year prospective cohort study, with biannual CD4 measurement. Follow-up was censored at the first severe morbidity event or at HAART initiation. Changes in CD4 values were studied by jointly modelling (a) the correlation between repeated measures through a linear mixed model and (b) the time to drop-out through a survival model.

RESULTS

690 patients were followed up during 1,382 person-years. Contrasting with the baseline CD4 count and percentage, which were associated with numerous variables, the slopes of both CD4 count and CD4 percentage in the absence of severe morbidity episode were only associated with the follow-up time and with the baseline body mass index (BMI). The mean annual natural decrease in CD4 count (CD4%) was estimated at -81/mm3 (-2.2%), -69/mm3 (-1.7%), and -55/mm3 (-1.2%) for patients with baseline BMI at 16 kg/m2, 20.4 kg/m2, and 25 kg/m2, respectively (p < .001). A steeper decline in the CD4 count was independently associated with a shorter event-free follow-up time.

CONCLUSION

These estimates of the CD4 natural decrease in sub-Saharan African patients, while they did not experience any episode of severe morbidity and before they initiate HAART, are in the bracket of those previously reported in industrialized countries. In sub-Saharan African settings with CD4 count being measured less frequently than in industrialized countries, the CD4 should be monitored more closely among adults with low BMI.

Long-term evolution of CD4 count in patients with a plasma HIV RNA persistently <500 copies/mL during treatment with antiretroviral drugs

BACKGROUND

The increase in CD4 count may reach a plateau after some duration of virological response to highly active antiretroviral therapy (HAART).

METHODS

A total of 1281 HIV-infected patients initiating HAART were enrolled in the AntiPROtease (APROCO) cohort. We investigated determinants of increase in CD4 count using longitudinal mixed models in patients who maintained a plasma HIV RNA <500 HIV-1 RNA copies/mL.

RESULTS

A total of 870 patients had a virological response at month 4. The median follow-up time was 57 months. Mean estimated increases in CD4 count in patients with persistent virological response were 29.9 cells/muL/month before month 4, 6.4 cells/microL/month between months 4 and 36, and 0.7 cells/microL/month (not significantly different from 0) after month 36. Three factors were associated with a significantly positive CD4 count slope after month 36: male gender (+0.9), no history of antiretroviral therapy at baseline (+1.7) and baseline CD4 count <100 cells/microL (+2.6). In patients who maintained a virological response after 5 years of HAART, a CD4 count >500 cells/microL was achieved in 83% of those with a baseline CD4 count >or=200 cells/microL and in 45% of those with a baseline CD4 count <200 cells/microL.

CONCLUSION

The increase in CD4 count reaches a plateau after 3 years of virological response. Even if patients initiating HAART with low CD4 counts still show a CD4 count increase after 3 years, it remains insufficient to overcome immune deficiency in all patients.

Trends and determinants of severe morbidity in HIV-infected patients: the ANRS CO3 Aquitaine Cohort, 2000-2004

OBJECTIVE

The aim of the study was to characterize the causes, trends and determinants of severe morbidity in a large cohort of HIV-infected patients between 2000 and 2004.

METHOD

Severe morbid events were defined as medical events associated with hospitalization or death. Epidemiological and biological data were recorded at the time of the morbid event. Trends were estimated using Poisson regression.

RESULTS

Among 3863 individuals followed between 2000 and 2004, 1186 experienced one or more severe events, resulting in 1854 hospitalizations or deaths. The severe events recorded included bacterial infections (21%), AIDS events (20%), psychiatric events (10%), cardiovascular events (9%), digestive events including cirrhosis (7%), viral infections (6%) and non-AIDS cancers (5%). Between 2000 and 2004, the incidence rate of AIDS events decreased from 60 to 20 per 1000 person-years, that of bacterial infections decreased from 45 to 24 per 1000 person-years, and that of psychiatric events decreased from 26 to 14 per 1000 person-years (all P<0.01), whereas the incidences of cardiovascular events and of non-AIDS cancers remained stable at 14 and 10 per 1000 person-years, on average, respectively.

CONCLUSION

Severe morbidity has shifted from AIDS-related to non-AIDS-related events during the course of HIV infection in developed countries. Limiting endpoints to AIDS events and death is insufficient to describe HIV disease progression in the era of combination antiretroviral therapy.

Practical identifiability of HIV dynamics models

We study the practical identifiability of parameters, i.e., the accuracy of the estimation that can be hoped, in a model of HIV dynamics based on a system of non-linear Ordinary Differential Equations (ODE). This depends on the available information such as the schedule of the measurements, the observed components, and the measurement precision. The number of patients is another way to increase it by introducing an appropriate statistical "population" framework. The impact of each improvement of the experimental condition is not known in advance but it can be evaluated via the Fisher Information Matrix (FIM). If the non-linearity of the biological model, as well as the complex statistical framework makes computation of the FIM challenging, we show that the particular structure of these models enables to compute it as precisely as wanted. In the HIV model, measuring HIV viral load and total CD4+ count were not enough to achieve identifiability of all the parameters involved. However, we show that an appropriate statistical approach together with the availability of additional markers such as infected cells or activated cells should considerably improve the identifiability and thus the usefulness of dynamical models of HIV.

Maximum likelihood estimation in dynamical models of HIV

The study of dynamical models of HIV infection, based on a system of nonlinear ordinary differential equations (ODE), has considerably improved the knowledge of its pathogenesis. While the first models used simplified ODE systems and analyzed each patient separately, recent works dealt with inference in non-simplified models borrowing strength from the whole sample. The complexity of these models leads to great difficulties for inference and only the Bayesian approach has been attempted by now. We propose a full likelihood inference, adapting a Newton-like algorithm for these particular models. We consider a relatively complex ODE model for HIV infection and a model for the observations including the issue of detection limits. We apply this approach to the analysis of a clinical trial of antiretroviral therapy (ALBI ANRS 070) and we show that the whole algorithm works well in a simulation study.

Analyse d'une variable quantitative censurée à gauche: exemple du taux de procalcitonine

BACKGROUND

When the sensitivity of an assay used to quantify a marker is poor, some of the values are below the detection limit resulting in left-censoring. Analysis of such data requires appropriate statistical techniques. In this study, we aimed at comparing various methods used to deal with left-censored outcome in regression analysis.

METHODS

The application was a real study evaluating the performance of procalcitonin for the diagnosis of bacterial infections among elderly patients. Among 85 patients, eleven had a procalcitonin value below the detection limit. A simulation study was then performed with data sampled according to a Gaussian distribution with parameters estimated on observed data. Various levels of left-censoring were simulated (13, 25 and 50%). A linear regression model was used to explain procalcitonin variations according to another marker, C reactive protein. To handle left-censoring, several methods were used: complete case analysis, simple imputation and multiple imputation methods, and parametric modelling. In the simulation study, estimations according to different methods were compared in terms of bias and mean square error according to each left-censoring level. Estimations obtained with real data were also compared according to the methods used. All analyses were implemented using SAS software.

RESULTS

In the simulation study, parametric modelling using maximum likelihood showed best performances whatever the level of censoring. On the other hand, methods using complete cases and simple imputation by the detection limit were highly skewed. On observed data, estimations of the slope varied slightly according to the methods. However the p-values (Wald test) of beta=0 varied from 0.0001 to 0.13 leading to different decisions according to the method used.

CONCLUSION

Left-censoring handling in data analysis requires special attention, as different methods may yield results leading to different conclusions.

Determinants of response to first HAART regimen in antiretroviral-naïve patients with an estimated time since HIV seroconversion

OBJECTIVE

To study the determinants of immunological and virological response to highly active antiretroviral therapy (HAART) in naïve patients, adjusting for time since HIV-1 seroconversion.

DESIGN

Data from HIV-cohort studies where dates of seroconversion have been reliably estimated. Methods In previously untreated patients, short- and long-term marker responses from HAART initiation (three or more antiretroviral drugs) to the end of follow-up or any treatment modification were considered using mixed effects models accounting for undetectable HIV viral load and informative dropout.

RESULTS

In total, 943 patients were treated with a first HAART regimen for a median of 29 months. In adjusted analyses, compared with a reference group of homosexual men without AIDS initiating treatment 4 years after seroconversion, injecting drug users (IDUs) were treated at similar CD4 and HIV RNA levels but had poorer short-term virological response (2.54 vs 2.13 log(10) HIV-1 RNA copies/mL at 1.5 months, P=0.03) and poorer long-term immunological response (522 vs 631 cells/microL at 24 months, P<0.0001). Although individuals with AIDS at HAART initiation had lower CD4 counts (206 vs 382 cells/microL, P<0.0001), their immunological responses were similar to those of individuals without AIDS. Similarly, individuals further from seroconversion started HAART at lower CD4 counts (e.g. 311 vs 382 cells/microL at vs before 9 years from seroconversion, P<0.0001), but had similar CD4 responses. However, they experienced poorer long-term virological response (0.67 log(10) copies/mL/year smaller decline, P<0.0001) compared to those treated before 9 years from seroconversion.

CONCLUSION

Taking into account the time elapsed since seroconversion, this study suggests that careful choices of initial treatment should be made and intensive follow-up carried out in high-risk subgroups such as IDUs who have poorer responses.

Carotid intima-media thickness is slightly increased over time in HIV-1-infected patients

OBJECTIVES

HIV-infected patients are at risk of atherosclerosis and cardiovascular diseases. In a 12-month follow-up study, we aimed to investigate changes in carotid intima-media thickness (IMT), a surrogate marker of atherosclerosis, and its determinants in HIV-1-infected patients.

METHODS

Our multicentre prospective longitudinal cohort study included 346 HIV-infected patients, for each of whom two IMT measurements were taken by B-mode ultrasonography at baseline (M0) and 1 year later (M12).

RESULTS

We observed a significant but moderate increase in the common carotid artery (CCA) median IMT, from 0.54 to 0.56 mm (P<10(-4)), i.e. an increase of 0.020 mm (95% confidence interval 0.012-0.029). There was a significant association between cross-sectional CCA IMT measures at M12 and conventional cardiovascular risk factors (higher CCA IMT with older age, P<10(-4); male gender, P=0.02; tobacco consumption, P=0.05), as well as higher CD4 cell count at M12 (>median 455 cells/microL, P=0.01). Only CD4 cell count at M0 was strongly and positively associated with the variation in IMT between M0 and M12 (P=4 x 10(-3)). IMT progression was +0.0020 mm for the lowest quartile of CD4 cell count distribution at M0, i.e. 3-253 cells/microL, +0.010 mm for 253-402 cells/microL, +0.043 mm for 402-590 cells/microL, and +0.028 mm for 590-2270 cells/microL. No association was found with type or duration of antiretroviral exposure.

CONCLUSIONS

Conventional cardiovascular risk factors are major determinants of IMT evolution. The link between immunological status and carotid IMT requires further study.

The use of the Framingham equation to predict myocardial infarctions in HIV-infected patients: comparison with observed events in the D:A:D Study

BACKGROUND

The D:A:D (Data Collection on Adverse Events of Anti-HIV Drugs) Study, a prospective observational study on a cohort of 23 468 patients with HIV infection, indicated that the incidence of myocardial infarction (MI) increased by 26% per year of exposure to combination antiretroviral treatment (CART). However, it remains unclear whether the observed increase in the rate of MI in this population can be attributed to changes in conventional cardiovascular risk factors.

OBJECTIVE

To compare the number of MIs observed among participants in the D:A:D Study with the number predicted by assuming that conventional cardiovascular risk equations apply to patients with HIV infection.

METHODS

The Framingham equation, a conventional cardiovascular risk algorithm, was applied to individual patient data in the D:A:D Study to predict rates of MI by duration of CART. A series of sensitivity analyses were performed to assess the effect of model and data assumptions. Predictions were extrapolated to provide 10-year risk estimates, and various scenarios were modelled to assess the expected effect of different interventions.

RESULTS

In patients receiving CART, the observed numbers of MIs during D:A:D follow up were similar to or somewhat higher than predicted numbers: 9 observed vs 5.5 events predicted, 14 vs 9.8, 22 vs 14.9, 31 vs 23.2 and 47 vs 37.0 for<1 year, 1-2 years, 2-3 years, 3-4 years and >4 years CART exposure, respectively. In patients who had not received CART, the observed number of MIs was fewer than predicted (3 observed vs 7.6 predicted). Nine per cent of the study population have a predicted 10-year risk of MI above 10%, a level usually associated with initiation of intervention on risk factors.

CONCLUSIONS

A consistent feature of all analyses was that observed and predicted rates of MI increased in a parallel fashion with increased CART duration, suggesting that the observed increase in risk of MI may at least in part be explained by CART-induced changes in conventional risk factors. These findings provide guidance in terms of choosing lifestyle or therapeutic interventions to decrease those risk factors in much the same way as in persons without HIV infection.

Determinants of clinical progression in antiretroviral-naive HIV-infected patients starting highly active antiretroviral therapy. Aquitaine Cohort, France, 1996-2002

OBJECTIVES

To determine the factors associated with clinical progression (AIDS events and death) in antiretroviral-naive patients who have begun highly active antiretroviral therapy (HAART).

METHODS

HIV-infected patients naive to antiretroviral therapy were included in a prospective hospital-based cohort who began HAART between June 1996 and December 2001. Progression was explained by baseline characteristics using Cox proportional hazards models.

RESULTS

Overall, data for 709 patients were analysed. In multivariate analysis, factors associated with an increased risk of progression were CD4 count < 50 cells/microL [hazard ratio (HR) = 13.0 (95% confidence interval 3.8-44.3)] and between 50 and 199 cells/microL [HR = 5.1 (1.6-16.3)], when compared with patients with CD4 count>350 cells/microL; AIDS events before HAART prescription [HR = 2.1 (1.2-3.7)]; CD8 count < 400 cells/microL [HR = 1.8 (1.1-3.0)]; and older age (HR = 1.2 by 10 years (1.0-1.5)]. In a second model including CD4 percentage, factors associated with progression were CD4 < 10% [HR = 6.3 (2.2-17.9)] and 10%<CD4 < 15% [HR = 4.2 (1.4-12.5)], when compared with patients with CD4 > 20%; CD8 count; AIDS events before HAART prescription; and older age. In a third model including the CD4:CD8 ratio, factors associated with progression were CD4:CD8 < 15% [HR = 8.2 (2.3-28.8)] and 15% < CD4:CD8 < 30% [HR = 4.6 (1.3-16.0)], when compared with patients with CD4:CD8 > 45%; AIDS events before HAART prescription; and older age. The Akaike information criteria for model analysis were 803, 805 and 815, respectively.

CONCLUSIONS

Consideration of CD4 level in terms of CD4:CD8 ratio or CD4 percentage can be a good alternative to absolute CD4 count. Other prognostic factors such as older age, CD8 count < 400 cells/microL and AIDS events also have to be considered in the decision to initiate HAART.

Modélisation conjointe de données longitudinales quantitatives et de délais censurés

BACKGROUND

In epidemiology, we are often interested in the association between the evolution of a quantitative variable and the onset of an event. The aim of this paper is to present a joint model for the analysis of Gaussian repeated data and survival time. Such models allow, for example, to perform survival analysis when a time-dependent explanatory variable is measured intermittently, or to study the evolution of a quantitative marker conditionally to an event.

METHODS

They are constructed by combining a mixed model for repeated Gaussian variables and a survival model which can be parametric or semi-parametric (Cox model).

RESULTS

We discuss the hypotheses underlying the different joint models proposed in the literature and the necessary assumptions for maximum likelihood estimation. The interest of these methods is illustrated with a study of the natural history of dementia in a cohort of elderly persons.

Fibrates or statins and lipid plasma levels in 245 patients treated with highly active antiretroviral therapy. Aquitaine Cohort, France, 1999-2001

OBJECTIVES

To describe the effect and tolerance of lipid-lowering drugs (LLD; fibrates and statins) in patients treated with highly active antiretroviral therapy (HAART).

METHODS

A prospective study was performed in a large cohort of HAART-treated HIV-infected patients using guidelines for case management of dyslipidaemia. Inclusion criteria were: age over 18 years; HIV infection treated with HAART including at least one protease inhibitor (PI) or nonnucleoside reverse transcriptase inhibitor (NNRTI); total cholesterol (TC)>5.5 mmol/L and/or triglycerides (TG)>2.2 mmol/L; and initiation of fibrates or statins. Viral load, CD4 cell count, plasma lipid levels and liver enzymes were measured at baseline (M0) and 3(M3) and 12(M12) months thereafter. Muscular enzymes were not assessed.

RESULTS

Fibrates were prescribed to 179 patients and statins to 66 patients. There was a significant mean decrease of TG in the fibrates group between M0 and M3 [-2.29 mmol/L; 95% confidence interval (CI)=-3.53, -1.05; P<10(-4)] and between M0 and M12 (-2.25 mmol/L; 95% CI:-4.23, -0.29; P<10(-4)); a significant decrease of TC was also noticed between M0 and M3 (-0.55 mmol/L; 95% CI:-0.95, -0.15; P=0.008) but not at M12 (-0.33 mmol/L; 95% CI:-0.94, 0.26; P=0.27). In the statins group, TC significantly decreased between M0 and M3 (-0.78 mmol/L; 95% CI:-1.3, -0.27; P=0.004) and between M0 and M12 (-0.70 mmol/L; 95% CI:-1.31, -0.09; P=0.03). There was no significant difference between the decreases of TG or TC observed in patients treated with PI or NNRTI. There were no significant modifications of viral load, CD4 cell count or liver enzymes during the 12 months of followup.

CONCLUSIONS

In this HAART-treated cohort, fibrates and statins were safe and associated with a favourable but moderate effect on lipid plasma levels.

[Epidemiology of atherosclerotic cardiovascular risk in HIV-1 infected patients]

OBVIOUS EFFICACY, BUT ALSO RISKS: The administration of combinations of highly active antiretroviral since 1996 has greatly decreased morbidity and mortality in HIV-1 infected patients. Side effects associated with these treatments were rapidly described. Other than abnormal distribution of fat, the dyslipidemia and insulin-resistance leads to a pro-atherogenic profile in these patients. Since 1998, the publication of several observations of acute cardio-vascular events in HIV-1 infected patients most often treated by antiretrovirals has raised the question of the increased risk of atherosclerosis in these patients. Today, some studies have been published and several are ongoing to evaluate the cardiovascular risk in HIV-1 infected patients. TWO FUNDAMENTAL QUESTIONS: The purpose of this review was to critically report the results of published studies in order to answer the two major questions: do HIV-1 infected patients have an increased cardiovascular risk compared to non infected subjects, and what are these cardiovascular risk factors? REPLY ELEMENTS: The review of available published results is in favour of an increased risk in HIV-1 infected patients compared to the general population. However, it is difficult to evaluate the importance of the absolute risk compared to the benefits of antiretroviral treatments and also the inherent effect of antiretroviral molecules. The main risk factors seem to be classical risk factors such as smoking habits, age and dyslipidemia that are partly due themselves to highly active antiretroviral treatments

A cohort study of nevirapine tolerance in clinical practice: French Aquitaine Cohort, 1997-1999

We performed a retrospective study to evaluate, under routine circumstances, the tolerance and immunovirological changes associated with antiretroviral regimens that contain nevirapine in 137 patients (88% were antiretroviral experienced). During a mean follow-up of 11 months, 33% of patients reported side effects attributed to nevirapine, and 21% discontinued treatment because of poor tolerance. Administration of antihistamines or corticosteroids at the initiation of treatment was not protective against adverse events (relative risk, 0.82; 95% confidence interval, 0.49-1.38). The proportion of patients with hepatitis C virus (HCV) and/or hepatitis B virus (HBV) coinfection who had alanine aminotransferase levels of >100 IU/L increased from 19.4% at baseline to 42.9% at month 12 of follow-up (P=.02). We noticed a significant increase of the proportion of patients with total cholesterol levels of >5.5 mM (P=.02). We have shown that there is a high level of discontinuation of nevirapine therapy in clinical practice and that side effects were not prevented by administration of antihistamines or corticosteroids. Coinfection with HCV or HBV increased the risk of hepatotoxicity, which lead to the cautious use of nevirapine for such patients.