Statin Use and Incidence of Parkinson's Disease in Women from the French E3N Cohort Study.

BACKGROUND: Statins represent candidates for drug repurposing in Parkinson's disease (PD). Few studies examined the role of reverse causation, statin subgroups, and dose-response relations based on time-varying exposures. OBJECTIVES: We examined whether statin use is associated with PD incidence while attempting to overcome the limitations described previously, especially reverse causation. METHOD: We used data from the E3N cohort study of French women (follow-up, 2004-2018). Incident PD was ascertained using multiple sources and validated by experts. New statin users were identified through linked drug claims. We set up a nested case-control study to describe trajectories of statin prescriptions and medical consultations before diagnosis. We used time-varying multivariable Cox proportional hazards regression models to examine the statins-PD association. Exposure indexes included ever use, cumulative duration/dose, and mean daily dose and were lagged by 5 years to address reverse causation. RESULTS: The case-control study (693 cases, 13,784 controls) showed differences in case-control trajectories, with changes in the 5 years before diagnosis in cases. Of 73,925 women (aged 54-79 years), 524 developed PD and 11,552 started using statins in lagged analyses. Ever use of any statin was not associated with PD (hazard ratio [HR] = 0.87, 95% confidence interval [CI] = 0.67-1.11). Alternatively, ever use of lipophilic statins was significantly associated with lower PD incidence (HR = 0.70, 95% CI = 0.51-0.98), with a dose-response relation for the mean daily dose (P-linear trend = 0.02). There was no association for hydrophilic statins. CONCLUSION: Use of lipophilic statins at least 5 years earlier was associated with reduced PD incidence in women, with a dose-response relation for the mean daily dose. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Comparison of cohort and nested case-control designs for estimating the effect of time-varying drug exposure on the risk of adverse event in the presence of ties.

Cohort and nested case-control (NCC) designs are frequently used in pharmacoepidemiology to assess the associations of drug exposure that can vary over time with the risk of an adverse event. Although it is typically expected that estimates from NCC analyses are similar to those from the full cohort analysis, with moderate loss of precision, only few studies have actually compared their respective performance for estimating the effects of time-varying exposures (TVE). We used simulations to compare the properties of the resulting estimators of these designs for both time-invariant exposure and TVE. We varied exposure prevalence, proportion of subjects experiencing the event, hazard ratio, and control-to-case ratio and considered matching on confounders. Using both designs, we also estimated the real-world associations of time-invariant ever use of menopausal hormone therapy (MHT) at baseline and updated, time-varying MHT use with breast cancer incidence. In all simulated scenarios, the cohort-based estimates had small relative bias and greater precision than the NCC design. NCC estimates displayed bias to the null that decreased with a greater number of controls per case. This bias markedly increased with higher proportion of events. Bias was seen with Breslow's and Efron's approximations for handling tied event times but was greatly reduced with the exact method or when NCC analyses were matched on confounders. When analyzing the MHT-breast cancer association, differences between the two designs were consistent with simulated data. Once ties were taken correctly into account, NCC estimates were very similar to those of the full cohort analysis.

Trends in psychotropic drug consumption among French military personnel during the COVID-19 epidemic.

BACKGROUND: The coronavirus disease (COVID-19) pandemic may have had significant mental health consequences for military personnel, which is a population already exposed to psychological stress. To assess the potential impact of the COVID-19 pandemic, we analyzed the dispensing of three classes of psychotropic drugs (anxiolytics, hypnotics, and antidepressants) among French military personnel. METHODS: A retrospective analysis was conducted using the individualized medico-administrative data of persons insured by the National Military Social Security Fund from the National Health Data System. All active French military personnel aged 18-64 who received outpatient care and to whom drugs were dispensed between January 1, 2019, and April 30, 2021, were included from the French national health database. Rate ratios of dispensed anxiolytics, hypnotics and antidepressants (based on drug reimbursement) were estimated from negative binomial regressions before and after the start of the COVID-19 pandemic. RESULTS: Three hundred eighty-one thousand seven hundred eleven individuals were included. Overall, 45,148 military personnel were reimbursed for anxiolytics, 10,637 for hypnotics, and 4328 for antidepressants. Drugs were dispensed at a higher rate in 2020 and 2021 than in 2019. There was a notable peak at the beginning of the first lockdown followed by a decrease limited to the duration of the first lockdown. During the first lockdown only, there were temporary phenomena including a brief increase in drug dispensing during the first week followed by a decrease during the rest of lockdown, possibly corresponding to a stocking-up effect. For the study period overall, while there was a significant downward trend in psychotropic drug dispensing before the occurrence of COVID-19 (p < 0.001), the pandemic period was associated with an increase in dispensed anxiolytics (rate ratio, 1.03; 95% CI, 1.02-1.04, p < 0.05), hypnotics (rate ratio, 1.13; 95% CI, 1.11-1.16, p < 0.001) and antidepressants (rate ratio, 1.12; 95% CI, 1.10-1.13, p < 0.001) in the military population. CONCLUSIONS: The COVID-19 pandemic has probably had a significant impact on the mental health of French military personnel, as suggested by the trends in dispensed psychotropic drugs. The implementation of mental health prevention measures should be investigated for this population.

Identifying Protective Drugs for Parkinson's Disease in Health-Care Databases Using Machine Learning.

BACKGROUND: Available treatments for Parkinson's disease (PD) are only partially or transiently effective. Identifying existing molecules that may present a therapeutic or preventive benefit for PD (drug repositioning) is thus of utmost interest. OBJECTIVE: We aimed at detecting potentially protective associations between marketed drugs and PD through a large-scale automated screening strategy. METHODS: We implemented a machine learning (ML) algorithm combining subsampling and lasso logistic regression in a case-control study nested in the French national health data system. Our study population comprised 40,760 incident PD patients identified by a validated algorithm during 2016 to 2018 and 176,395 controls of similar age, sex, and region of residence, all followed since 2006. Drug exposure was defined at the chemical subgroup level, then at the substance level of the Anatomical Therapeutic Chemical (ATC) classification considering the frequency of prescriptions over a 2-year period starting 10 years before the index date to limit reverse causation bias. Sensitivity analyses were conducted using a more specific definition of PD status. RESULTS: Six drug subgroups were detected by our algorithm among the 374 screened. Sulfonamide diuretics (ATC-C03CA), in particular furosemide (C03CA01), showed the most robust signal. Other signals included adrenergics in combination with anticholinergics (R03AL) and insulins and analogues (A10AD). CONCLUSIONS: We identified several signals that deserve to be confirmed in large studies with appropriate consideration of the potential for reverse causation. Our results illustrate the value of ML-based signal detection algorithms for identifying drugs inversely associated with PD risk in health-care databases. © 2022 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Analyzing cohort studies with interval-censored data: A new model-based linear rank-type test.

To compare two or more survival distributions with interval-censored data, various nonparametric tests have been proposed. Some are based on the G ρ $$ {G}^{\rho } $$ -family introduced by Harrington and Fleming (1991) that allows flexibility for situations in which the hazard ratio decreases monotonically to unity. However, it is unclear how to choose the appropriate value of the parameter ρ $$ \rho $$ . In this work, we propose a novel linear rank-type test for analyzing interval-censored data that derived from a proportional reversed hazard model. We show its relationship with decreasing hazard ratio. This test statistic provides an alternative to the G ρ $$ {G}^{\rho } $$ -based test statistics by bypassing the choice of the ρ $$ \rho $$ parameter. Simulation results show its good behavior. Two studies on breast cancer and drug users illustrate its practical uses and highlight findings that would have been overlooked if other tests had been used. The test is easy to implement with standard software and can be used for a wide range of situations with interval-censored data to test the equality of survival distributions between two or more independent groups.

Hospitalizations for Stroke in France During the COVID-19 Pandemic Before, During, and After the National Lockdown.

BACKGROUND AND PURPOSE: In France, the entire population was put under a total lockdown from March 17 to May 11, 2020 during the peak of the coronavirus disease 2019 (COVID-19) pandemic. Whether the lockdown had consequences on the management of medical emergencies such as stroke and transient ischemic attack (TIA) has yet to be fully evaluated. This article describes hospitalization rates for acute stroke in 2 French regions that experienced contrasting rates of COVID-19 infection, before, during, and after the nationwide lockdown (January to June 2020). METHODS: All patients admitted for acute stroke/TIA into all public and private hospitals of the 2 study regions were included. Data were retrieved from the National Hospitalization Database (PMSI). In the most affected region (Grand-Est), the hospitalization rates observed in April 2020 were compared with the rates in the same period in the least affected region (Occitanie) and in the 3 prior years (2017-2019). RESULTS: There was a significant decline in hospitalization rates for stroke/TIA within the region most affected by COVID-19 during the month of April 2020 compared with previous years, while no significant change was seen in the least affected region. After lockdown, we observed a fast rebound in the rate of hospitalization for stroke/TIA in the most affected region, contrasting with a slower rebound in the least affected region. In both regions, patients with COVID-19 stroke more frequently had ischemic stroke, a nonsignificant greater prevalence of diabetes, they were less frequently admitted to stroke units, and mortality was higher than in patients without COVID-19. CONCLUSIONS: Our results demonstrates a significant drop in stroke/TIA hospitalizations and a fast recovery after the end of the French lockdown in the most affected region, while the least affected region saw a nonsignificant drop in stroke/TIA hospitalizations and a slow recovery. These results and recommendations could be used by the health authorities to prepare for future challenges.

Chronic respiratory diseases are predictors of severe outcome in COVID-19 hospitalised patients: a nationwide study.

BACKGROUND: Influenza epidemics were initially considered to be a suitable model for the COVID-19 epidemic, but there is a lack of data concerning patients with chronic respiratory diseases (CRDs), who were supposed to be at risk of severe forms of COVID-19. METHODS: This nationwide retrospective cohort study describes patients with prior lung disease hospitalised for COVID-19 (March-April 2020) or influenza (2018-2019 influenza outbreak). We compared the resulting pulmonary complications, need for intensive care and in-hospital mortality depending on respiratory history and virus. RESULTS: In the 89 530 COVID-19 cases, 16.03% had at least one CRD, which was significantly less frequently than in the 45 819 seasonal influenza patients. Patients suffering from chronic respiratory failure, chronic obstructive pulmonary disease, asthma, cystic fibrosis and pulmonary hypertension were under-represented, contrary to those with lung cancer, sleep apnoea, emphysema and interstitial lung diseases. COVID-19 patients with CRDs developed significantly more ventilator-associated pneumonia and pulmonary embolism than influenza patients. They needed intensive care significantly more often and had a higher mortality rate (except for asthma) when compared with patients with COVID-19 but without CRDs or patients with influenza. CONCLUSIONS: Patients with prior respiratory diseases were globally less likely to be hospitalised for COVID-19 than for influenza, but were at higher risk of developing severe COVID-19 and had a higher mortality rate compared with influenza patients and patients without a history of respiratory illness.

Impact of COVID-19 and lockdowns on pulmonary embolism in hospitalized patients in France: a nationwide study.

BACKGROUND: This study assessed the impact of the COVID-19 epidemic on overall hospitalizations for pulmonary embolism (PE) in France in comparison with previous years, and by COVID-19 and non-COVID-19 status. METHODS: Hospitalization data (2017-2020) were extracted from the French National Discharge database (all public and private hospitals). We included all patients older than 18 years hospitalized during the 3 years and extracted PE status and COVID-19 status (from March 2020). Age, sex and risk factors for PE (such as obesity, cancer) were identified. We also extracted transfer to an intensive care unit (ICU) and hospital death. The number of PE and the frequency of death in patients in 2019 and 2020 were described by month and by COVID-19 status. Logistic regressions were performed to identify the role of COVID-19 among other risk factors for PE in hospitalized patients. RESULTS: The overall number of patients hospitalized with PE increased by about 16% in 2020 compared with 2019, and mortality also increased to 10.3% (+ 1.2%). These increases were mostly linked to COVID-19 waves, which were associated with PE hospitalization in COVID-19 patients (PE frequency was 3.7%; 2.8% in non-ICU and 8.8% in ICU). The final PE odds ratio for COVID-19 hospitalized patients was 4 compared with other hospitalized patients in 2020. The analyses of PE in non-COVID-19 patients showed a 2.7% increase in 2020 compared with the previous three years. CONCLUSION: In 2020, the overall number of patients hospitalized with PE in France increased compared to the previous three years despite a considerable decrease in scheduled hospitalizations. Nevertheless, proactive public policy focused on the prevention of PE in all patients should be encouraged.

New adaptive lasso approaches for variable selection in automated pharmacovigilance signal detection.

BACKGROUND: Adverse effects of drugs are often identified after market introduction. Post-marketing pharmacovigilance aims to detect them as early as possible and relies on spontaneous reporting systems collecting suspicious cases. Signal detection tools have been developed to mine these large databases and counts of reports are analysed with disproportionality methods. To address disproportionality method biases, recent methods apply to individual observations taking into account all exposures for the same patient. In particular, the logistic lasso provides an efficient variable selection framework, yet the choice of the regularization parameter is a challenging issue and the lasso variable selection may give inconsistent results. METHODS: We propose a new signal detection methodology based on the adaptive lasso. We derived two new adaptive weights from (i) a lasso regression using the Bayesian Information Criterion (BIC), and (ii) the class-imbalanced subsampling lasso (CISL), an extension of stability selection. The BIC is used in the adaptive lasso stage for variable selection. We performed an extensive simulation study and an application to real data, where we compared our methods to the existing adaptive lasso, and recent detection approaches based on lasso regression or propensity scores in high dimension. For both studies, we evaluate the methods in terms of false discoveries and sensitivity. RESULTS: In the simulations and the application, both proposed adaptive weights show equivalent or better performances than the other competitors, with an advantage for the CISL-based adaptive weights. CISL and lasso regression using BIC are solid alternatives. CONCLUSION: Our proposed adaptive lasso is an appealing methodology for signal detection in pharmacovigilance. Although we cannot rely on test theory, our approaches show a low and stable False Discovery Rate in all simulation settings. All methods evaluated in this work are implemented in the adapt4pv R package.

Risk of Intracranial Aneurysm and Dissection and Fluoroquinolone Use: A Case-Time-Control Study.

Background and Purpose- Fluoroquinolone use is associated with an increased risk of aortic aneurysm and dissection. We investigated this risk of arterial wall injury on intracranial arteries, given the similar pathophysiological mechanisms for aneurysm and dissection in both types of arteries. Methods- A case-time-control study was conducted using French National Insurance databases covering >60 million inhabitants. Cases were aged ≥18 years with first ruptured intracranial aneurysm and dissection between 2010 and 2015. For each case, fluoroquinolone use was compared between the exposure-risk window (day 30-day 1 before the outcome) and matched control windows (day 120-day 91, day 150-day 121, and day 180-day 151) and adjusted for time-varying confounders; potential time-trend for exposure was controlled using an age- and sex-matched reference group. Amoxicillin use was studied similarly for indication bias controlling. The potential excess of risk conveyed by fluoroquinolones was assessed by the ratio of OR for fluoroquinolones to that for amoxicillin. Results- Of the 7443 identified cases, 75 had been exposed to fluoroquinolones in the prior 180 days, including 16 in the 30-day at-risk window (385/97 cases exposed to amoxicillin, respectively). The adjusted OR for fluoroquinolones was 1.26 (95%CI, 0.65-2.41) and that for amoxicillin of 1.36 (95% CI, 1.05-1.78). Ratio of OR for fluoroquinolones to that for amoxicillin was estimated at 0.92 (95% CI, 0.46-1.86). Result was similar when extending outcome definition to unruptured events (ratio of OR for fluoroquinolones to that for amoxicillin, 0.97 [95% CI, 0.61-1.53]). Conclusions- This study did not evidence an excess of risk of intracranial aneurysm or dissection with fluoroquinolone use.

Accounting for indirect protection in the benefit-risk ratio estimation of rotavirus vaccination in children under the age of 5 years, France, 2018.

Background Rotavirus is a major cause of severe gastroenteritis in children worldwide. The disease burden has been substantially reduced in countries where rotavirus vaccines are used. Given the risk of vaccine-induced intussusception, the benefit­risk balance of rotavirus vaccination has been assessed in several countries, however mostly without considering indirect protection effects. Aim We performed a benefit­risk analysis of rotavirus vaccination accounting for indirect protection in France among the 2018 population of children under the age of 5 years. Methods To incorporate indirect protection effects in the benefit formula, we adopted a pseudo-vaccine approach involving mathematical approximation and used a simulation design to provide uncertainty intervals. We derived background incidence distributions from quasi-exhaustive health claim data. We examined different coverage levels and assumptions regarding the waning effects and intussusception case fatality rate. Results With the current vaccination coverage of < 10%, the indirect effectiveness was estimated at 6.4% (+/− 0.4). For each hospitalisation for intussusception, 277.0 (95% uncertainty interval: (165.0­462.1)) hospitalisations for rotavirus gastroenteritis were prevented. Should 90% of infants be vaccinated, indirect effectiveness would reach 57.9% (+/− 3.7) and the benefit­risk ratio would be 192.4 (95% uncertainty interval: 116.4­321.3). At a coverage level of 50%, indirect protection accounted for 27% of the prevented rotavirus gastroenteritis cases. The balance remained in favour of the vaccine even in a scenario with a high assumption for intussusception case fatality. Conclusions These findings contribute to a better assessment of the rotavirus vaccine benefit­risk balance.

Instrumental variable analysis in the context of dichotomous outcome and exposure with a numerical experiment in pharmacoepidemiology.

BACKGROUND: In pharmacoepidemiology, the prescription preference-based instrumental variables (IV) are often used with linear models to solve the endogeneity due to unobserved confounders even when the outcome and the endogenous treatment are dichotomous variables. Using this instrumental variable, we proceed by Monte-Carlo simulations to compare the IV-based generalized method of moment (IV-GMM) and the two-stage residual inclusion (2SRI) method in this context. METHODS: We established the formula allowing us to compute the instrument's strength and the confounding level in the context of logistic regression models. We then varied the instrument's strength and the confounding level to cover a large range of scenarios in the simulation study. We also explore two prescription preference-based instruments. RESULTS: We found that the 2SRI is less biased than the other methods and yields satisfactory confidence intervals. The proportion of previous patients of the same physician who were prescribed the treatment of interest displayed a good performance as a proxy of the physician's preference instrument. CONCLUSIONS: This work shows that when analysing real data with dichotomous outcome and exposure, appropriate 2SRI estimation could be used in presence of unmeasured confounding.

Prescription drug use during pregnancy in France: a study from the national health insurance permanent sample.

PURPOSE: To provide an up-to-date account of drug prescription during pregnancy in France from 2011 to 2014 using the permanent sample of the French national computerized healthcare database and with a focus on recommended supplementations, fetotoxic drugs and teratogenic drugs. METHODS: All pregnancies identified by the International Classification of Diseases, 10th Revision codes list in the hospitalization database, lasting more than 9 weeks of amenorrhea and whose delivery occurred between 01/01/2011 and 12/31/2014, were included. Drugs delivered between the trimester before and until the end of the pregnancy were included. Drug exposure prevalence was calculated for each year and according to pregnancy trimesters. RESULTS: The study included 28,491 pregnancies with a median number of 9 [5-13] (median [IQ range]) drugs delivered. The most prescribed drug class was antianemia (in 72.5% of exposed). The prescription rate of recommended vitamins (B9 and D) increased over the study period (+10%). Influenza vaccination also increased but remained at a low rate (1%). Exposure to fetotoxic drugs decreased as pregnancy advanced. Exposure to the main teratogenic antiepileptics was stable over the study period. Low-income pregnant women had a higher average drug consumption except for recommended vitamins. CONCLUSION: Pregnant French women are among the largest consumers of prescription medications worldwide. Overall, the dispensation trends observed in this study are in line with the recommendations of the French National College of Gynecologists and Obstetricians. Nevertheless, while being low, exposure to fetotoxic drugs, teratogenic drugs or those under safety alerts still occurred. Supplementations and vaccines in low-income pregnant women should also be increased.

A benefit-risk analysis of rotavirus vaccination, France, 2015.

IntroductionTwo vaccines available for protection against rotavirus gastroenteritis (RVGE), Rotarix and RotaTeq, have contributed to a large decrease in the incidence of paediatric diarrhoea in countries where they have been used. However, they have also led to a small increase in the risk of intussusception. Methods: We compare the number of prevented hospitalisations for RVGE to the number of vaccine-induced hospitalised intussusceptions in France. Results: With 9.5% coverage (French 2015 estimation), vaccination was estimated to prevent, annually, a median of 1,074 hospitalisations (2.5th and 97.5th percentiles (2.5th-97.5th): 810-1,378) and 1.4 deaths (2.5th-97.5th: 1.2-1.6) from RVGE. It was also estimated to cause, annually, 5.0 hospitalisations (2.5th-97.5th: 3.2-7.7) and 0.005 deaths (2.5th-97.5th: 0.001-0.015) from intussusception. The benefit-risk ratio is therefore 214 (2.5th-97.5th: 128-362) for hospitalisations and 273 (2.5th-97.5th: 89-1,228) for deaths. Under a hypothetical 92% coverage, rotavirus vaccination with Rotarix would avoid 10,459 (2.5th-97.5th: 7,702-13,498) hospitalisations for RVGE and induce 47.0 (2.5th-97.5th: 25.1-81.4) hospitalisations for intussusception annually, thereby preventing 13.7 (2.5th-97.5th: 11.1-15.2) deaths and inducing 0.05 (2.5th-97.5th: 0.01-0.15) deaths. Conclusion: The benefit-risk ratio in France is similar to that of other European countries.

Postmenopausal Hormone Therapy and Risk of Stroke: Impact of the Route of Estrogen Administration and Type of Progestogen.

BACKGROUND AND PURPOSE: The benefit/risk analysis of hormone therapy in postmenopausal women is not straightforward and depends on cardiovascular disease. Evidence supports the safety of transdermal estrogens and the importance of progestogens for thrombotic risk. However, the differential association of oral and transdermal estrogens with stroke remains poorly investigated. Furthermore, there are no data regarding the impact of progestogens. METHODS: We set up a nested case-control study of ischemic stroke (IS) within all French women aged 51 to 62 years between 2009 and 2011 without personal history of cardiovascular disease or contraindication to hormone therapy. Participants were identified using the French National Health Insurance database, which includes complete drug claims for the past 3 years and French National hospital data. We identified 3144 hospitalized IS cases who were matched for age and zip code to 12 158 controls. Conditional logistic regression was used to estimate odds ratios (OR) and 95% confidence intervals (95% CI). RESULTS: Compared with nonusers, the adjusted ORs of IS were1.58 (95% CI, 1.01-2.49) in oral estrogen users and 0.83 (0.56-1.24) in transdermal estrogens users (P<0.01). There was no association of IS with use of progesterone (OR, 0.78; 95% CI, 0.49-1.26), pregnanes (OR, 1.00; 95% CI, 0.60-1.67), and nortestosterones (OR, 1.26; 95% CI, 0.62-2.58), whereas norpregnanes increased IS risk (OR, 2.25; 95% CI, 1.05-4.81). CONCLUSIONS: Both route of estrogen administration and progestogens were important determinants of IS. Our findings suggest that transdermal estrogens might be the safest option for short-term hormone therapy use.

Bayesian model selection in logistic regression for the detection of adverse drug reactions.

Spontaneous adverse event reports have a high potential for detecting adverse drug reactions. However, due to their dimension, the analysis of such databases requires statistical methods. In this context, disproportionality measures can be used. Their main idea is to project the data onto contingency tables in order to measure the strength of associations between drugs and adverse events. However, due to the data projection, these methods are sensitive to the problem of coprescriptions and masking effects. Recently, logistic regressions have been used with a Lasso type penalty to perform the detection of associations between drugs and adverse events. On different examples, this approach limits the drawbacks of the disproportionality methods, but the choice of the penalty value is open to criticism while it strongly influences the results. In this paper, we propose to use a logistic regression whose sparsity is viewed as a model selection challenge. Since the model space is huge, a Metropolis-Hastings algorithm carries out the model selection by maximizing the BIC criterion. Thus, we avoid the calibration of penalty or threshold. During our application on the French pharmacovigilance database, the proposed method is compared to well-established approaches on a reference dataset, and obtains better rates of positive and negative controls. However, many signals (i.e., specific drug-event associations) are not detected by the proposed method. So, we conclude that this method should be used in parallel to existing measures in pharmacovigilance. Code implementing the proposed method is available at the following url: https://github.com/masedki/MHTrajectoryR.

Dynamics of the risk of smoking-induced lung cancer: a compartmental hidden Markov model for longitudinal analysis.

BACKGROUND: To account for the dynamic aspects of carcinogenesis, we propose a compartmental hidden Markov model in which each person is healthy, asymptomatically affected, diagnosed, or deceased. Our model is illustrated using the example of smoking-induced lung cancer. METHODS: The model was fitted on a case-control study nested in the European Prospective Investigation into Cancer and Nutrition study, including 757 incident cases and 1524 matched controls. Estimation was done through a Markov Chain Monte Carlo algorithm, and simulations based on the posterior estimates of the parameters were used to provide measures of model fit. We performed sensitivity analyses to assess robustness of our findings. RESULTS: After adjusting for its impact on exposure duration, age was not found to independently drive the risk of lung carcinogenesis, whereas age at starting smoking in ever-smokers and time since cessation in former smokers were found to be influential. Our data did not support an age-dependent time to diagnosis. The estimated time between onset of malignancy and clinical diagnosis ranged from 2 to 4 years. Our approach yielded good performance in reconstructing individual trajectories in both cases (sensitivity >90%) and controls (sensitivity >80%). CONCLUSION: Our compartmental model enabled us to identify time-varying predictors of risk and provided us with insights into the dynamics of smoking-induced lung carcinogenesis. Its flexible and general formulation enables the future incorporation of disease states, as measured by intermediate markers, into the modeling of the natural history of cancer, suggesting a large range of applications in chronic disease epidemiology.

Matching methods to create paired survival data based on an exposure occurring over time: a simulation study with application to breast cancer.

BACKGROUND: Paired survival data are often used in clinical research to assess the prognostic effect of an exposure. Matching generates correlated censored data expecting that the paired subjects just differ from the exposure. Creating pairs when the exposure is an event occurring over time could be tricky. We applied a commonly used method, Method 1, which creates pairs a posteriori and propose an alternative method, Method 2, which creates pairs in "real-time". We used two semi-parametric models devoted to correlated censored data to estimate the average effect of the exposure HR¯(t): the Holt and Prentice (HP), and the Lee Wei and Amato (LWA) models. Contrary to the HP, the LWA allowed adjustment for the matching covariates (LWAa) and for an interaction (LWAi) between exposure and covariates (assimilated to prognostic profiles). The aim of our study was to compare the performances of each model according to the two matching methods. METHODS: Extensive simulations were conducted. We simulated cohort data sets on which we applied the two matching methods, the HP and the LWA. We used our conclusions to assess the prognostic effect of subsequent pregnancy after treatment for breast cancer in a female cohort treated and followed up in eight french hospitals. RESULTS: In terms of bias and RMSE, Method 2 performed better than Method 1 in designing the pairs, and LWAa was the best model for all the situations except when there was an interaction between exposure and covariates, for which LWAi was more appropriate. On our real data set, we found opposite effects of pregnancy according to the six prognostic profiles, but none were statistically significant. We probably lacked statistical power or reached the limits of our approach. The pairs' censoring options chosen for combination Method 2 - LWA had to be compared with others. CONCLUSIONS: Correlated censored data designing by Method 2 seemed to be the most pertinent method to create pairs, when the criterion, which characterized the pair, was an exposure occurring over time. In such a setting, the LWA was the most appropriate model.

Estimating time-to-onset of adverse drug reactions from spontaneous reporting databases.

BACKGROUND: Analyzing time-to-onset of adverse drug reactions from treatment exposure contributes to meeting pharmacovigilance objectives, i.e. identification and prevention. Post-marketing data are available from reporting systems. Times-to-onset from such databases are right-truncated because some patients who were exposed to the drug and who will eventually develop the adverse drug reaction may do it after the time of analysis and thus are not included in the data. Acknowledgment of the developments adapted to right-truncated data is not widespread and these methods have never been used in pharmacovigilance. We assess the use of appropriate methods as well as the consequences of not taking right truncation into account (naive approach) on parametric maximum likelihood estimation of time-to-onset distribution. METHODS: Both approaches, naive or taking right truncation into account, were compared with a simulation study. We used twelve scenarios for the exponential distribution and twenty-four for the Weibull and log-logistic distributions. These scenarios are defined by a set of parameters: the parameters of the time-to-onset distribution, the probability of this distribution falling within an observable values interval and the sample size. An application to reported lymphoma after anti TNF- α treatment from the French pharmacovigilance is presented. RESULTS: The simulation study shows that the bias and the mean squared error might in some instances be unacceptably large when right truncation is not considered while the truncation-based estimator shows always better and often satisfactory performances and the gap may be large. For the real dataset, the estimated expected time-to-onset leads to a minimum difference of 58 weeks between both approaches, which is not negligible. This difference is obtained for the Weibull model, under which the estimated probability of this distribution falling within an observable values interval is not far from 1. CONCLUSIONS: It is necessary to take right truncation into account for estimating time-to-onset of adverse drug reactions from spontaneous reporting databases.

Increased risk of narcolepsy in children and adults after pandemic H1N1 vaccination in France.

An increased incidence of narcolepsy in children was detected in Scandinavian countries where pandemic H1N1 influenza ASO3-adjuvanted vaccine was used. A campaign of vaccination against pandemic H1N1 influenza was implemented in France using both ASO3-adjuvanted and non-adjuvanted vaccines. As part of a study considering all-type narcolepsy, we investigated the association between H1N1 vaccination and narcolepsy with cataplexy in children and adults compared with matched controls; and compared the phenotype of narcolepsy with cataplexy according to exposure to the H1N1 vaccination. Patients with narcolepsy-cataplexy were included from 14 expert centres in France. Date of diagnosis constituted the index date. Validation of cases was performed by independent experts using the Brighton collaboration criteria. Up to four controls were individually matched to cases according to age, gender and geographic location. A structured telephone interview was performed to collect information on medical history, past infections and vaccinations. Eighty-five cases with narcolepsy-cataplexy were included; 23 being further excluded regarding eligibility criteria. Of the 62 eligible cases, 59 (64% males, 57.6% children) could be matched with 135 control subjects. H1N1 vaccination was associated with narcolepsy-cataplexy with an odds ratio of 6.5 (2.1-19.9) in subjects aged<18 years, and 4.7 (1.6-13.9) in those aged 18 and over. Sensitivity analyses considering date of referral for diagnosis or the date of onset of symptoms as the index date gave similar results, as did analyses focusing only on exposure to ASO3-adjuvanted vaccine. Slight differences were found when comparing cases with narcolepsy-cataplexy exposed to H1N1 vaccination (n=32; mostly AS03-adjuvanted vaccine, n=28) to non-exposed cases (n=30), including shorter delay of diagnosis and a higher number of sleep onset rapid eye movement periods for exposed cases. No difference was found regarding history of infections. In this sub-analysis, H1N1 vaccination was strongly associated with an increased risk of narcolepsy-cataplexy in both children and adults in France. Even if, as in every observational study, the possibility that some biases participated in the association cannot be completely ruled out, the associations appeared robust to sensitivity analyses, and a specific analysis focusing on ASO3-adjuvanted vaccine found similar increase.