Search

Background The Marginal Structural Cox Model (Cox-MSM), an alternative approach to handle time-dependent confounder, was introduced for survival analysis and applied to estimate the joint causal effect of two time-dependent nonrandomized treatments on survival among HIV-positive subjects. Nevertheless, Cox-MSM performance in the case of multiple treatments has not been fully explored under different degree of time-dependent confounding for treatments or in case of interaction between treatments. We aimed to evaluate and compare the performance of the marginal structural Cox model (Cox-MSM) to the standard Cox model in estimating the treatment effect in the case of multiple treatments under different scenarios of time-dependent confounding and when an interaction between treatment effects is present. Methods We specified a Cox-MSM with two treatments including an interaction term for situations where an adverse event might be caused by two treatments taken simultaneously but not by each treatment taken alone. We simulated longitudinal data with two treatments and a time-dependent confounder affected by one or the two treatments. To fit the Cox-MSM, we used the inverse probability weighting method. We illustrated the method to evaluate the specific effect of protease inhibitors combined (or not) to other antiretroviral medications on the anal cancer risk in HIV-infected individuals, with CD4 cell count as time-dependent confounder. Results Overall, Cox-MSM performed better than the standard Cox model. Furthermore, we showed that estimates were unbiased when an interaction term was included in the model. Conclusion Cox-MSM may be used for accurately estimating causal individual and joined treatment effects from a combination therapy in presence of time-dependent confounding provided that an interaction term is estimated. Electronic supplementary material The online version of this article (10.1186/s12874-017-0434-1) contains supplementary material, which is available to authorized users.

Background. Calendar trends in virologic failure (VF) among human immunodeficiency virus (HIV)-infected patients can help to evaluate the performance of healthcare systems and the need for new antiretroviral therapy (ART). We examined the time trend in the rate of VF beyond 6 months of ART between 1997 and 2011 in France. Methods. We included patients from the French Hospital Database on HIV who received at least 6 months of ART. VF was defined as 2 consecutive plasma HIV-RNA values >500 copies/mL or as 1 value >500 copies/mL followed byatreatment switch. We adjusted for patients’ characteristics by fitting a multivariable generalized estimating equation logistic regression model with an exchangeable covariance matrix. Results. A total of 81738 patients were enrolled, and median follow-up was 112.4 months. Median CD4 count was 333 cells/µL, and 23% of patients had HIV infection classified as Centers for Disease Control and Prevention stage C. Overall, 29.3% of patients received single/dual-drug ART initially, and 45.4% of patients experienced at least 1 episode of VF during follow-up. The percentage of patients with VF fell from 61.5% in 1997–1998 to 9.7% in 2009–2011 (P

International audience; The French Hospital Database on HIV (FHDH) is a hospital-based multicentre open cohort with inclusions ongoing since 1989. The research objectives focus mainly on mid- and long-term clinical outcomes and therapeutic strategies, as well as severe AIDS and non-AIDS morbidities, and public health issues relative to HIV infection. FHDH also serves to describe HIV-infected patients receiving hospital care in France. FHDH includes data on more than 120,000 HIV-infected patients from 70 French general or university hospitals distributed throughout France. Patients are eligible for inclusion if they are infected by HIV-1 or HIV-2 and give their written informed consent. Standardized variables are collected at each outpatient visit or hospital admission during which a new clinical manifestation is diagnosed, a new treatment is prescribed or a change in biological markers is noted, and/or at least every 6 months. Since its inception, variables collected in FHDH include demographic characteristics, HIV-related biological markers, the date and type of AIDS and non AIDS-defining events, antiretroviral treatments and the date and causes of death, as reported in the medical records. Since 2005, data have also been collected on: co-infection with hepatitis B or C virus; alcohol and tobacco use; and non HIV-related biomarkers. Anyone can submit a research project by completing a standardized form available on the FHDH website (http://www.ccde.fr/_fold/fl-1385734776-429.pdf) or from the corresponding author, describing the context and objectives of the study. All projects are reviewed by the scientific committee.

International audience; Background - Chronic kidney disease (CKD) is frequent in individuals infected with human immunodeficiency virus (HIV). Progression to end-stage renal disease can be slowed by appropriate medical management. Methods - To assess whether active promotion of guidelines improves CKD management, we conducted a cluster randomized controlled trial within the French Hospital Database on HIV (FHDH-ANRS CO4). We randomized 46 centers participating in the FHDH to either simple information on guideline availability or active promotion with a multifaceted and repeated intervention comprising reminders and audit feedback and targeting of local opinion leaders carried out between April 2009 and April 2010. Outcome measure was CKD management adequacy assessed before and 2 years after the beginning of the intervention in HIV-infected patients with moderate to severe CKD. CKD management was considered adequate in case of referral to a nephrologist or if proteinuria, blood pressure, low-density lipoprotein cholesterol level, and glycemia had been measured during the previous year and medications had been prescribed when necessary. Results - Three hundred six patients were enrolled, of whom 238 (78%) completed the 2 years of follow-up. During the study period, the percentage of patients receiving adequate CKD management improved from 64.1% to 70.4% (+6.3%) in the active arm and from 68.3% to 75.6% (+7.3%) in the control arm (adjusted mean difference, -0.7 percentage points [95% confidence interval: -9.2 to 7.9]; P = .95). The biggest impact of active promotion was on the management of proteinuria and blood pressure. Conclusions - Adequate compliance with CKD management guidelines improved slightly between 2009 and 2011, with no difference between the simple information and active promotion arms. Clinical trials registration - CCTIRS 10.150 and CNIL DR-2010-379.

Background Recent studies have shown a decrease in the incidence of herpes zoster (HZ) among human immunodeficiency virus (HIV)-infected patients since the combined antiretroviral therapy (cART) era, but more data are needed on a possible increase in the risk early after cART initiation. Methods We studied HZ incidence and risk factors among patients followed in the French Hospital Database on HIV (FHDH) between 1992 and 2011. Standardized incidence ratios (SIRs) were used for comparison with the general population between 2005 and 2008. The risk of HZ following cART initiation (0-5 and ≥6 months) was studied with Poisson regression models. Results A total of 7167 cases of incident HZ were diagnosed among 91 044 individuals (583 125 person-years [PY]). The incidence declined significantly, from 2955 per 100 000 PY in 1992-1996 to 628 per 100 000 PY in 2009-2011. This decline was mainly explained by cART (relative risk [RR], 0.60; 95% confidence interval {CI}, .57-.64). The risk of HZ was associated with low CD4 cell counts, high HIV RNA levels, low CD4/CD8 ratios, and prior AIDS. Compared to the general population, the risk of HZ was higher in HIV-infected patients (overall SIR, 2.7; 95% CI, 2.6-2.9), particularly those aged 15-44 years (SIR, 4-6). In ART-naive patients, a moderate increase in the HZ risk was observed during the first 6 months of cART, with a peak at 3 months (RR, 1.47; 95% CI, 1.26-1.73), a finding that disappeared after adjustment for the current CD4 cell count (RR, 1.03; 95% CI, .81-1.32). Conclusions The risk of HZ has declined markedly among HIV-infected patients in the cART era, but remains 3 times higher than in the general population. The risk increases moderately during the first 6 months of cART.

International audience; Using the sample of 107 families with at least two asthmatic siblings, as part of the EGEA study, we have investigated linkage to asthma (or atopy) and genetic heterogeneity according to the presence/absence of atopy (or asthma) using two approaches: (1) the triangle test statistic (TTS), which considers the identical by descent (IBD) distribution among affected sib-pairs discordant for another associated phenotype (eg asthmatic sib-pairs discordant for atopy) and (2) the predivided sample test (PST), which compares the IBD distribution of marker alleles between affected sib-pairs concordant and discordant for the associated phenotype. Two regions, 8p and 12q, already reported to be linked to both asthma and atopy, were examined here. A total of 20 asthmatic sib-pairs discordant for atopy and 24 atopic pairs discordant for asthma were analyzed by both TTS and PST methods and 83 pairs with atopic asthma by PST. Some evidence for linkage was observed for two markers in the 8p23.3-p23.2 region; D8S504 for asthma with genetic heterogeneity according to the presence/absence of atopy and D8S503 for atopy with genetic heterogeneity according to the presence/absence of asthma. In the 12q14.2-q21.33 region, there was also some evidence of linkage to two markers, D12S83 and D12S95, for atopy and asthma, respectively, with genetic heterogeneity according to the presence/absence of the associated trait. Provided the small distance between the two markers on either 8p (16 cM) or 12q (21 cM), it is unclear whether one or two genetic factors are involved in either region.

The aim of this study was to compare, under different models of gene-environment (G x E) interaction, the power to detect linkage and G x E interaction of different tests using affected sib-pairs. Methods considered were: 1) the maximum likelihood lod-score (MLS), based on the distribution of parental alleles identical by descent (IBD) in affected sibs; 2) the sum of the MLS (sMLS) calculated in affected sib-pairs with 2, 1, or 0 sibs exposed; 3) the predivided sample test (PST), which compares the IBD distribution between affected sib-pairs with 2, 1, or 0 sibs exposed; 4) the triangle test statistic (TTS), which uses the IBD distribution among discordant affected sib-pairs (one exposed, one unexposed); and 5) the mean interaction test (MIT), based on the regression of the proportion of alleles shared IBD among affected sib-pairs on the exposure among sib-pairs. The MLS, sMLS, and MIT allow detection of linkage. However, the sMLS and MIT account for a possible G x E interaction without testing it. In contrast, the PST and the TTS allow detection of both linkage and G x E interaction. Results showed that when exposure cancels the effect of the gene, or changes the direction of this effect (i.e., the protective allele becomes the risk allele), the PST, sMLS, and MIT may provide, under some models, greater power to detect linkage than the MLS. Under models where exposure changes the direction of the effect of the gene, the TTS test may also be more powerful than the other tests accounting for G x E interaction. Under the other models, the MLS remains the most powerful test to detect linkage. However, only the PST and TTS allow the detection of G x E interaction.

Despite tremendous efforts, few genes involved in the susceptibility for complex disorders have been identified. One explanation is that these disorders are a result of an interaction between genes and environment, and under such conditions, it may be difficult to measure the true genetic effect without accounting for the interaction. Umbach and Weinberg ([2000] Am. J. Hum. Genet. 66:251–261) proposed an association test which looks at the joint effects of genotype and environment, using case-parent trios. In this study, we explore under which conditions accounting for G×E interaction enhances one's ability to detect the role of genetic factors in complex diseases. Using asymptotic power calculations, we investigate the power to detect the gene effect over varying exposure frequencies and different scenarios of G×E interaction. We show that for a given sample size, interaction scenario, and allele frequency, the actual gain in power while accounting for the interaction depends on the magnitude of the exposure frequency: the largest gains are seen for relatively low exposure frequencies. Moreover, a loss of power can be observed when the exposure is frequent and/or the exposure effect is strong. If we consider a gene with a disease allele frequency of 0.2, with no effect in the absence of exposure, an exposure with a 10-fold increase risk and a G×E relative risk of 2, then when the exposure frequency is 0.1, accounting for G×E interaction increases the power to detect the gene effect in 200 trios by 10%; alternatively, when the exposure frequency is 0.9, it decreases the power by 15%. Genet Epidemiol 24:200–207, 2003. © 2003 Wiley-Liss, Inc.

Purpose To study recent trends in the incidence of anal cancer in HIV-infected patients receiving long-term combined antiretroviral treatment (cART) compared with the general population. Patients and Methods From the French Hospital Database on HIV, we identified 263 cases of invasive anal squamous cell carcinoma confirmed histologically between 1992 and 2008. We compared incidence rates of anal cancer across four calendar periods: 1992-1996 (pre-cART period), 1997-2000 (early cART period), and 2001-2004 and 2005-2008 (recent cART periods). Standardized incidence ratios (SIRs) were calculated by using general population incidence data from the French Network of Cancer Registries. Results In HIV-infected patients, the hazard ratio (HR) in the cART periods versus the pre-cART period was 2.5 (95% CI, 1.28 to 4.98). No difference was observed across the cART calendar periods (HR, 0.9; 95% CI, 0.6 to 1.3). In 2005-2008, HIV-infected patients compared with the general population had an excess risk of anal cancer, with SIRs of 109.8 (95% CI, 84.6 to 140.3), 49.2 (95% CI, 33.2 to 70.3), and 13.1 (95% CI, 6.8 to 22.8) for men who have sex with men (MSM), other men, and women, respectively. Among patients with CD4 cell counts above 500/μL for at least 2 years, SIRs were 67.5 (95% CI, 41.2 to 104.3) when the CD4 nadir was less than 200/μL for more than 2 years and 24.5 (95% CI, 17.1 to 34.1) when the CD4 nadir was more than 200/μL. Conclusion Relative to that in the general population, the risk of anal cancer in HIV-infected patients is still extremely high, even in patients with high current CD4 cell counts. cART appears to have no preventive effect on anal cancer, particularly in MSM.

International audience; OBJECTIVES: To compare virological effectiveness in patients who continued on a virologically successful first-line boosted protease inhibitor (PI)-containing combination antiretroviral therapy (cART) regimen or who switched to a PI-free cART including efavirenz, nevirapine or abacavir. METHODS: From the French Hospital Database on HIV, we selected 439 patients with undetectable viral load (VL) on a first-line boosted PI-containing cART regimen who switched to a PI-free combination including efavirenz, nevirapine or abacavir. Each of these patients was matched with three patients who continued to take their first-line cART regimen, on the basis of gender, age, CD4 cell count, VL, date of cART initiation and the duration of VL undetectability. Time to virological failure (VF) was analysed with Kaplan-Meier curves and Cox models. RESULTS: The 12 month probabilities of VF were 3.7% and 5.7% in non-switch and switch patients, respectively, and 3.9%, 7.2% and 9.0% in patients switching to efavirenz-, nevirapine- and abacavir-containing cART, respectively. After adjustment, only patients switching to abacavir-containing cART had a higher risk of VF than non-switch patients (adjusted hazard ratio, 1.99; 95% confidence interval, 1.05-3.79). CONCLUSIONS: Switching from a virologically successful first-line boosted PI-containing cART regimen to a non-nucleoside reverse transcriptase inhibitor-containing cART regimen containing either efavirenz or nevirapine is virologically safe, while switching to abacavir-containing cART should be avoided.

OBJECTIVE: To estimate the prevalence and characteristics of long-term nonprogressor (LTNP) and HIV controller patients in a very large French cohort of HIV1-infected patients. METHODS: In the French Hospital Database on HIV [FHDH, Agence Nationale de Recherches sur le SIDA et les hépatites virales (ANRS) CO4], we selected patients who had been seen in 2005, who had been infected for more than 8 years, who were treatment-naive, and who remained asymptomatic. Patients with these characteristics then categorized as follows: LTNP (> or =8 years of HIV infection and CD4 cell nadir > or =500/microl), elite LTNP (> or =8 years of HIV infection, CD4 cell nadir > or =600/microl, and a positive CD4 slope), HIV controllers (>10 years of HIV infection with 90% of plasma viral load values < or =500 copies/ml), and elite controllers (same as HIV controllers, but with last plasma viral load value < or =50 copies/ml in 2005). RESULTS: Among the 46 880 HIV1-infected patients followed in 2005 in the French Hospital Database on HIV, 0.4% (N = 202) were LTNP, 0.05% (N = 25) were elite LTNP, 0.22% (N = 101) were HIV controllers, and 0.15% (N = 69) were elite controllers. Ten elite LTNP patients (40%) were also HIV controllers, eight (32%) were elite controllers, and 60% had detectable plasma viral load (>50 copies/ml). Among the elite controllers, 32 (46%) were LTNP, eight (12%) were elite LTNP, and one-quarter had a last CD4 cell count less than 500/microl. CONCLUSION: LTNP, elite LTNP, HIV controller, and elite controller patients are rare phenotypes. Elite LTNP patients are less frequent than HIV controllers. There is little overlap among the four subgroups of patients.

Objective: To describe the cases of anal cancer that appeared in the French Hospital Database on HIV between 1992 and 2004 and to study risk factors of anal cancer. Methods: We examined the incidence rates of anal cancer between 1992 and 2004 and the risk associated among 86322 HIV-infected patients included in the French Hospital Database on HIV. Results: We identified 132 cases of anal cancer, including 124 cases in men (94%), of whom 75% had sex with men. Median age at diagnosis was 42.8 years (interquartile range: 36.9-49.4). At diagnosis, 103 patients (78%) were receiving combination antiretroviral therapy for a median of 37.1 months (interquartile range: 4.5-59.8). Median survival after anal cancer diagnosis was 5 years. The respective overall incidence rates of anal cancer per 100000 person-years between 1992 and March 1996, April 1996 to 1998 and between 1999 and 2004 were 11 (95% confidence interval, 4-17), 18 (95% confidence interval, 10-27) and 40 (95% confidence interval, 32-47). The risk of anal cancer was higher among men who have sex with men. After adjustment for age at inclusion in the study, as well as gender, the HIV transmission group, the nadir CD4 cell count and AIDS status, the incidence was higher in the years 1999-2004 than in between 1992 to March 1996 (hazard ratio, 2.5; 95% confidence interval, 1.2-5.3), with no change in the years 1999-2004. Conclusion: The incidence of anal cancer has increased among HIV-infected patients in France since 1996. Although an ascertainment bias cannot be excluded, data indicate that combination antiretroviral therapy does not prevent anal cancer in these patients. This supports the urgent need for developing anal cancer screening programs for HIV-infected men who have sex with men.

Background: Contrary to current HIV/AIDS management guidelines, and despite the arrival of potent combination antiretroviral therapy (cART) many years ago, some patients are still treated with dual nucleoside reverse transcriptase inhibitor (NRTI) regimens. Methods: We selected 5222 patients who received dual NRTI therapy for at least 6 months during 1998 to 2002, representing 9.9% of the 52,981 ARV-treated patients recorded in the French Hospital Database on HIV Factors associated with switching to cART or with ARV discontinuation were identified by using Cox models. Results: The 3-year probabilities of switching to cART and of antiretroviral (ARV) drug discontinuation were 55.2% (95% confidence interval [CI]: 53.8 to 56.7) and 10.9% (95% CI: 10.1 to 11.8), respectively, whereas 1591 patients (30.5%) kept the dual NRTI therapy during all the study period. Place of birth and region of care did not influence the choice of treatment strategy. After adjustment, the likelihood of switching to cART was lower among women, intravenous drug users, and patients with an undetectable plasma viral load (pVL) on at least 1 occasion during follow-up; in contrast, it was higher among patients with AIDS and those with a low CD4 cell count at enrollment or at the last follow-up visit. The likelihood of ARV discontinuation was higher among women and intravenous drug users and lower among patients with a low CD4 cell count at inclusion or at the last follow-up visit and among patients with an undetectable pVL on at least 1 occasion during follow-up. The likelihood of switching to cART or discontinuing ARV drugs was higher among patients receiving zidovudine/zalcitabine or didanosine/ stavudine than among those receiving zidovudine/lamivudine. Conclusions: In France, until recent years, some patients (mainly women and intravenous drug users) were still receiving dual NRTI therapy despite free access to care and to highly effective ARV regimens. Dual NRTI therapy is gradually being replaced by cART, although some patients with satisfactory immunovirologic status are discontinuing all ARV drugs.

Despite tremendous efforts, few genes involved in the susceptibility for complex disorders have been identified. One explanation is that these disorders are a result of an interaction between genes and environment, and under such conditions, it may be difficult to measure the true genetic effect without accounting for the interaction. Umbach and Weinberg ([2000] Am. J. Hum. Genet. 66:251-261) proposed an association test which looks at the joint effects of genotype and environment, using case-parent trios. In this study, we explore under which conditions accounting for GxE interaction enhances one's ability to detect the role of genetic factors in complex diseases. Using asymptotic power calculations, we investigate the power to detect the gene effect over varying exposure frequencies and different scenarios of GxE interaction. We show that for a given sample size, interaction scenario, and allele frequency, the actual gain in power while accounting for the interaction depends on the magnitude of the exposure frequency: the largest gains are seen for relatively low exposure frequencies. Moreover, a loss of power can be observed when the exposure is frequent and/or the exposure effect is strong. If we consider a gene with a disease allele frequency of 0.2, with no effect in the absence of exposure, an exposure with a 10-fold increase risk and a GxE relative risk of 2, then when the exposure frequency is 0.1, accounting for GxE interaction increases the power to detect the gene effect in 200 trios by 10%; alternatively, when the exposure frequency is 0.9, it decreases the power by 15%.

Three different samples of families with asthmatic patients, the German data set and the CSGA data set subdivided into Caucasian and African American groups, were analyzed using the maximum lod score statistic. Although different scores were obtained in each sample, the Wald's likelihood homogeneity test did not reveal any significant genetic heterogeneity. This may be due to the very large variance of the model-free linkage statistics.

We investigated the independent contributions of a candidate gene and an environmental factor, and the presence of gene x environment (G x E) interaction, in the etiology of a disease in the Genetic Analysis Workshop (GAW) 12 problem 2 simulated data using a two-stage approach utilizing both case-control and case-parent study designs. Using the case-control design, several SNPs within candidate gene 1 (CG1) and environmental factor 1 (dichotomized using the 75th percentile as a cut-off) (EXP) were independently associated with disease status, in models adjusted for age and sex. We found evidence of gene x environment (G x E) interaction between EXP and two single-nucleotide polymorphisms (SNPs) within CG1 using the case-control design. Using the case-parent study design in the same population, we detected association between SNPs within CG1 and disease, but no G x E interaction was detected.

It is generally believed that an early age at the onset of disease is associated with a stronger genetic component. Our aim here was to investigate both linkage and genetic heterogeneity of asthma, the latter corresponding to different genotype relative risks of a putative linked gene according to age at onset of asthma. This analysis was conducted in 107 French EGEA families with at least two asthmatic siblings, considering 157 markers that were part of our previous genome screen, using the TTS (the Triangle Test Statistic) which has been developed to detect both linkage and intra-sibpair genetic heterogeneity. This test has been applied to 38 asthmatic sib-pairs discordant for age at the onset of asthma. To confirm the existence of genetic heterogeneity, we also used the predivided sample test (PST) which compares the IBD (identity by descent) distribution of marker alleles between asthmatic sib-pairs concordant (67) and discordant (38) for the age at onset. The cutoff point used for the age at onset was 4 years, the median age at onset in our sample of asthmatic sibs. Linkage and genetic heterogeneity for a region located on chromosome 7q (at 109 cM from pter) were indicated by both tests, TTS (P=0.005, P>0.5 after correction for multiple testing) and PST (P=0.0001, 0.015 after correction). These results suggest a genetic factor on 7q involved in asthma with genotype relative risks differing according to age at onset of disease.

Crohn's disease (CD) is a complex genetic disorder for which a susceptibility gene, IBD1, has been mapped within the pericentromeric region of chromosome 16. In order to refine the location of IBD1, 77 multiplex CD families were genotyped for 26 microsatellite markers evenly spaced by approximately 1 cM. Nonparametric linkage analyses exhibited a maximum NPL score of 3.49 (P=2.37 ╫ 10-4) in a region centred by markers D16S3136, D16S3117 and D16S770. Simulation studies showed that the probability for IBD1 to be located in a 5 cM region around these markers was 70%. A 2.5 Mb YAC and BAC contig map spanning this genetic region on chromosome band 16q12 was built. TDT analyses demonstrated suggestive association between the 207 bp allele of D16S3136 (P

Coeliac disease (CD) is a malabsorptive disorder of the small intestine resulting from ingestion of gluten. The HLA risk factors involved in CD are well known but do not explain the whole genetic susceptibility. Several regions of potential linkage on chromosomes 3q, 5q, 10q, 11q, 15q and 19q have already been reported in the literature. These six regions were analyzed with the Maximum Lod Score method on a dense set of markers. A new sample of 89 Italian sibpairs was available for study. There was no evidence for linkage for any of the regions tested, except for chromosome 5q. For this region, our data, as well as a sample of 93 sibpairs from our first genome screen (Greco et al. 1998), are compatible with the presence of a risk factor for CD with a moderate effect.

International audience; The proportions of affected sibs sharing 2, 1 or 0 identical by descent parental marker alleles have been shown to conform to the 'triangle constraints' (Suarez, 1978; Holmans, 1993). It has also been shown (Dudoit & Speed, 1999) that the constraints are verified provided certain assumptions hold. In this study we explore a realistic situation in which the constraints fail due to the presence of a factor in which the sibs differ, a factor on which penetrance depends. This factor may be a characteristic of the trait (severe vs. mild form), or the presence/absence of an associated trait or an environmental factor. We show that under such situations, using the triangle constraints may lead to important loss of power to detect linkage by the MLS test. We propose here an alternative approach in order to detect both linkage and heterogeneity.