Olivier Bouaziz

Broad-Spectrum Antibiotic Treatment and Subsequent Childhood Type 1 Diabetes: A Nationwide Danish Cohort Study

Background Studies link antibiotic treatment and delivery by cesarean section with increased risk of chronic diseases through changes of the gut-microbiota. We aimed to evaluate the association of broad-spectrum antibiotic treatment during the first two years of life with subsequent onset of childhood type 1 diabetes and the potential effect-modification by mode of delivery. Materials and Methods A Danish nationwide cohort study including all singletons born during 1997–2010. End of follow-up by December 2012. Four national registers provided information on antibiotic redemptions, outcome and confounders. Redemptions of antibiotic prescriptions during the first two years of life was classified into narrow-spectrum or broad-spectrum antibiotics. Children were followed from age two to fourteen, both inclusive. The risk of type 1 diabetes with onset before the age of 15 years was assessed by Cox regression. A total of 858,201 singletons contributed 5,906,069 person-years, during which 1,503 children developed type 1 diabetes. Results Redemption of broad-spectrum antibiotics during the first two years of life was associated with an increased rate of type 1 diabetes during the following 13 years of life (HR 1.13; 95% CI 1.02 to 1.25), however, the rate was modified by mode of delivery. Broad-spectrum antibiotics were associated with an increased rate of type 1 diabetes in children delivered by either intrapartum cesarean section (HR 1.70; 95% CI 1.15 to 2.51) or prelabor cesarean section (HR 1.63; 95% CI 1.11 to 2.39), but not in vaginally delivered children. Number needed to harm was 433 and 562, respectively. The association with broad-spectrum antibiotics was not modified by parity, genetic predisposition or maternal redemption of antibiotics during pregnancy or lactation. Conclusions Redemption of broad-spectrum antibiotics during infancy is associated with an increased risk of childhood type 1 diabetes in children delivered by cesarean section.

Conditional density estimation in a censored single-index regression model

Under a single-index regression assumption, we introduce a new semiparametric procedure to estimate a conditional density of a censored response. The regression model can be seen as a generalization of Cox regression model and also as a profitable tool to perform dimension reduction under censoring. This technique extends the results of Delecroix et al. (2003). We derive consistency and asymptotic normality of our estimator of the index parameter by proving its asymptotic equivalence with the (uncomputable) maximum likelihood estimator, using martingales results for counting processes and arguments of empirical processes theory. Furthermore, we provide a new adaptive procedure which allows us both to chose the smoothing parameter involved in our approach and to circumvent the weak performances of Kaplan-Meier estimator (1958) in the right-tail of the distribution. Through a simulation study, we study the behavior of our estimator for small samples.

Semiparametric inference for the recurrent events process by means of a single-index model

In this paper, we introduce new parametric and semiparametric regression techniques for a recurrent event process subject to random right censoring. We develop models for the cumulative mean function and provide asymptotically normal estimators. Our semiparametric model which relies on a single-index assumption can be seen as a dimension reduction technique that, contrary to a fully nonparametric approach, is not stroke by the curse of dimensionality when the number of covariates is high. We discuss data-driven techniques to choose the parameters involved in the estimation procedures and provide a simulation study to support our theoretical results.

A penalized algorithm for event-specific rate models for recurrent events

We introduce a covariate-specific total variation penalty in two semiparametric models for the rate function of recurrent event process. The two models are a stratified Cox model, introduced in Prentice and others (1981. On the regression analysis of multivariate failure time data. Biometrika 68: , 373-379.), and a stratified Aalens additive model. We show the consistency and asymptotic normality of our penalized estimators. We demonstrate, through a simulation study, that our estimators outperform classical estimators for small-to-moderate sample sizes. Finally, an application to the bladder tumor data of Byar (1980. The veterans administration study of chemoprophylaxis for recurrent stage 1 bladder tumors: comparison of placebo, pyridoxine, and topical thiotepa. In Pavone-Macaluso, M. Smith, P. H. and F. Edsmyn (editors), Bladder Tumors and Others Topics in Urological Oncology, pp. 363-370.) is presented.

A Change-Point Model for Detecting Heterogeneity in Ordered Survival Responses

In this article, we suggest a new statistical approach considering survival heterogeneity as a breakpoint model in an ordered sequence of time-to-event variables. The survival responses need to be ordered according to a numerical covariate. Our estimation method will aim at detecting heterogeneity that could arise through the ordering covariate. We formally introduce our model as a constrained Hidden Markov Model, where the hidden states are the unknown segmentation (breakpoint locations) and the observed states are the survival responses. We derive an efficient Expectation-Maximization framework for maximizing the likelihood of this model for a wide range of baseline hazard forms (parametrics or nonparametric). The posterior distribution of the breakpoints is also derived, and the selection of the number of segments using penalized likelihood criterion is discussed. The performance of our survival breakpoint model is finally illustrated on a diabetes dataset where the observed survival times are ordered according to the calendar time of disease onset.

Is Placental Malaria a Long-term Risk Factor for Mild Malaria Attack in Infancy? Revisiting a Paradigm

Background Children born to mothers with placental malaria (PM) have been described as more susceptible to the occurrence of a first malaria infection. However, whether or not these children remain more at risk during infancy has never been explored. We aimed to determine if children born to mothers with PM are more susceptible to malaria and remain at higher risk between birth and 18 months. Methods Five hundred fifty children were followed up weekly with control of temperature and, if >37.5°C, both a rapid diagnostic test for malaria and a thick blood smear were performed. Taking into account environmental risk of infection, the relationship between occurrences of malaria attacks from birth to 18 months was modeled using Cox models for recurrent events. Results PM is not associated with an overall susceptibility to malaria but only with the delay of occurrence of the first malaria attack. Children born from mothers with PM tend to have an increased risk for the first malaria attack (hazard ratio [HR] = 1.33; P = .048) but not for subsequent ones (HR = 0.9; P = .46). Children who experienced 1 malaria attack were strongly at risk to develop subsequent infections independent of placental infection and environmental exposure. Conclusions These results are consistent with the existence of an individual susceptibility to malaria unrelated to PM. From a public health point of view, protecting children born to infected placenta remains a priority, but seems insufficient to account for other frail children for whom a biomarker of frailty needs to be found.

Computing Individual Risks Based on Family History in Genetic Disease in the Presence of Competing Risks

When considering a genetic disease with variable age at onset (e.g., familial amyloid neuropathy, cancers), computing the individual risk of the disease based on family history (FH) is of critical interest for both clinicians and patients. Such a risk is very challenging to compute because (1) the genotype X of the individual of interest is in general unknown, (2) the posterior distribution ℙ(X∣FH, T > t) changes with t (T is the age at disease onset for the targeted individual), and (3) the competing risk of death is not negligible. In this work, we present modeling of this problem using a Bayesian network mixed with (right-censored) survival outcomes where hazard rates only depend on the genotype of each individual. We explain how belief propagation can be used to obtain posterior distribution of genotypes given the FH and how to obtain a time-dependent posterior hazard rate for any individual in the pedigree. Finally, we use this posterior hazard rate to compute individual risk, with or without the competing risk of death. Our method is illustrated using the Claus-Easton model for breast cancer. The competing risk of death is derived from the national French registry.