Dirk Enders

Comparative risk of death in older adults treated with antipsychotics: A population-based cohort study

Although the use of antipsychotics has been associated with an increased risk of death, data on the safety of individual substances is scarce. We thus aimed to compare the risk of death in new users of individual antipsychotics aged =>65 years and conducted a cohort study in the German Pharmacoepidemiological Research Database between 2005 and 2011. Patients were followed from initiation of treatment until death, 90 days after cohort entry, end of insurance or the end of the study period. Multivariable cox regression was used to estimate confounder adjusted hazard ratios (aHR) of death for 14 individual antipsychotics compared to risperidone. In sensitivity analyses, we also applied high-dimensional propensity score (HDPS) methods to explore possible unmeasured confounding. In a cohort of 137,713 new users of antipsychotics, a higher risk of death was found for haloperidol (aHR: 1.45; 95% confidence interval: 1.35-1.55), levomepromazine (aHR: 1.34; 1.16-1.54), zuclopenthixol (aHR: 1.32; 1.02-1.72) and to a lesser extent for melperone (aHR: 1.13; 1.07-1.19) compared to risperidone. Lower risks were observed for quetiapine, prothipendyl, olanzapine, tiapride, clozapine, perazine and flupentixol. In subgroup analyses, levomepromazine and chlorprothixene were only associated with a higher risk of death in patients aged =>80 years and with dementia. The application of HDPS methods did not substantially change the results. In conclusion, our study suggests that initiation of haloperidol, levomepromazine, zuclopenthixol and chlorprothixene treatment is associated with an increased risk of death compared to risperidone and should be avoided in older patients except in palliative care when treatment alternatives are available.

The contribution of comorbidities to mortality in hospitalized patients with heart failure

BackgroundHeart failure (HF) with reduced ejection fraction (HFrEF) has a worse prognosis than HF with preserved EF (HFpEF). The study aimed to evaluate whether different comorbidity profiles of HFrEF- and HFpEF-patients or HF-specific mechanisms contribute to a greater extent to this difference.MethodsWe linked data from two health insurances to data from a cardiology clinic hospital information system. Patients with a hospitalization with HF in 2005–2011, categorized as HFrEF (EF < 45%) or HFpEF (EF ≥ 45%), were propensity score (PS) matched to controls without HF on comorbidites and medication to assure similar comorbidity profiles of patients and their respective controls. The balance of the covariates in patients and controls was compared via the standardized difference (SDiff). Age-standardized 1-year mortality rates (MR) with 95% confidence intervals (CI) were calculated.Results777 HFrEF-patients (1135 HFpEF-patients) were PS-matched to 3446 (4832) controls. Balance between patients and controls was largely achieved with a SDiff < 0.1 on most variables considered. The age-standardized 1-year MRs per 1000 persons in HFrEF-patients and controls were 267.8 (95% CI 175.9–359.8) and 86.1 (95% CI 70.0–102.3). MRs in HFpEF-patients and controls were 166.2 (95% CI 101.5–230.9) and 61.5 (95% CI 52.9–70.1). Thus, differences in MRs between patients and their controls were higher for HFrEF (181.7) than for HFpEF (104.7).ConclusionsGiven the similar comorbidity profiles between HF-patients and controls, the higher difference in mortality rates between HFrEF-patients and controls points more to HF-specific mechanisms for these patients, whereas for HFpEF-patients a higher contribution of comorbidity is suggested by our results.

Incidence of herpes zoster amongst adults varies by severity of immunosuppression

OBJECTIVES We examined the incidence of herpes zoster in immunocompromised adults (≥18 years) with different severities of immunosuppression and assessed the prevalence of complications and of various kinds of healthcare resource utilisation. METHODS German claims data from more than ten million adults were used to calculate annual incidence rates of herpes zoster for the years 2006-2012 and to analyse the prevalence of complications, physician visits, hospitalisations, and antiviral and analgesic treatments using a cohort design. The analyses were stratified by age, sex, and severity of immunosuppression, defined by immunocompromising conditions and drug therapies. RESULTS The incidence rate per 1000 person-years of herpes zoster was almost twice as high in immunocompromised patients (11.5 (95% confidence interval (CI): 11.4-11.6)) compared to immunocompetent subjects (5.9 (95% CI: 5.8-5.9)). The incidence rate was higher in highly immunocompromised patients (13.4 (95% CI: 13.2-13.6)) than in patients with a low severity of immunosuppression (10.0 (95% CI: 9.8-10.1)). These differences were observed for both sexes and in all age groups. Complications, outpatient physician visits, hospitalisations, and analgesic treatments occurred more frequently in immunocompromised patients as well. CONCLUSIONS Our results show that immunocompromised individuals are affected by the disease in particular and that the burden of herpes zoster is highest in severely immunocompromised patients.

Comparative risk for cardiovascular diseases of dipeptidyl peptidase-4 inhibitors vs. sulfonylureas in combination with metformin: Results of a two-phase study

AIMS The aim was to assess whether the use of additional data from the Disease Management Program (DMP) diabetes mellitus type 2 to minimize the potential for residual confounding will alter the estimated risk of either myocardial infarction, ischemic stroke or heart failure in patients with type 2 diabetes using sulfonylureas compared to dipeptidyl peptidase-4 (DPP-4) inhibitors in addition to metformin based on routine health care data. METHODS We conducted a nested two-phase case-control study using claims data of one German health insurance from 2004 to 2013 (phase 1) and data of the DMP from 2010 to 2013 (phase 2). Adjusted odds ratios (ORs) for the combined cardiovascular event myocardial infarction, ischemic stroke or heart failure were calculated using a two-phase logistic regression. RESULTS Phase 1 comprised 3179 patients (289 cases; 2890 controls) and phase 2 comprised 1968 patients (168 cases; 1800 controls). We observed an adjusted OR of 0.83 for the combined cardiovascular event (95% CI: 0.61-1.13). CONCLUSIONS We observed a non-significantly reduced risk for cardiovascular diseases in patients using DPP-4 inhibitors compared to sulfonylureas in addition to metformin. This finding was not altered by the inclusion of additional information of the DMP in the analysis. However, due to the low power of this study, further studies are needed to reproduce our findings.

The Potential of High‐Dimensional Propensity Scores in Health Services Research: An Exemplary Study on the Quality of Care for Elective Percutaneous Coronary Interventions

OBJECTIVE Evaluating the potential of the high-dimensional propensity score (HDPS) to control for residual confounding in studies analyzing quality of care based on administrative health insurance data. DATA SOURCE Secondary data from 2004 to 2009 from three German statutory health insurance providers. STUDY DESIGN We conducted a retrospective cohort study in patients with elective percutaneous coronary interventions (PCIs) and compared the mortality risk between the in- and outpatient setting using Cox regression. Adjustment for predefined confounders was performed using conventional propensity score (PS) techniques. Further, an HDPS was calculated based on predefined and empirically selected confounders from the database. PRINCIPAL FINDINGS Conventional PS methods showed a decreased mortality risk for outpatient compared to inpatient PCIs, while trimming of patients with nonoverlap in the HDPS distribution and weighting resulted in a comparable risk. Most comorbidities were less prevalent in the HDPS-trimmed population compared to the original one. CONCLUSION The HDPS methodology may reduce residual confounding by rendering the studied cohort more comparable through restriction. However, results cannot be generalized for the entire study population. To provide unbiased results, full assessment of all unmeasured confounders from proxy information in the database would be necessary.

Prevalence, Duration and Severity of Parkinson's Disease in Germany: A Combined Meta-Analysis from Literature Data and Outpatient Samples

Background: Epidemiological data on the prevalence of Parkinsons disease (PD) in Germany are limited. The aims of this study were to estimate the age- and gender-specific prevalence of PD in Germany as well as the severity and illness duration. Summary: A systematic literature search was performed in 5 different databases. European studies were included if they reported age- and gender-specific numbers of prevalence rates of PD. Meta-analytic approaches were applied to derive age- and gender-specific pooled prevalence estimates. Data of 4 German outpatient samples were incorporated to calculate the proportion of patients with PD in Germany grouped by Hoehn and Yahr (HY) stages and disease duration. In the German population, 178,169 cases of PD were estimated (prevalence: 217.22/100,000). The estimated relative illness duration was 40% with less than 5 years, 31% with 5-9 years, and 29% with more than 9 years. The proportions for different HY stages were estimated at 13% (I), 30% (II), 35% (III), 17% (IV), and 4% (V), respectively. Key Message: We provide an up-to-date estimation of age- and gender-specific as well as severity-based prevalence figures for PD in Germany. Further community studies are needed to estimate population-based severity distributions and distributions of non-motor symptoms in PD.

Robust versus consistent variance estimators in marginal structural Cox models: Variance estimation in Cox MSMs

In survival analyses, inverse-probability-of-treatment (IPT) and inverse-probability-of-censoring (IPC) weighted estimators of parameters in marginal structural Cox models are often used to estimate treatment effects in the presence of time-dependent confounding and censoring. In most applications, a robust variance estimator of the IPT and IPC weighted estimator is calculated leading to conservative confidence intervals. This estimator assumes that the weights are known rather than estimated from the data. Although a consistent estimator of the asymptotic variance of the IPT and IPC weighted estimator is generally available, applications and thus information on the performance of the consistent estimator are lacking. Reasons might be a cumbersome implementation in statistical software, which is further complicated by missing details on the variance formula. In this paper, we therefore provide a detailed derivation of the variance of the asymptotic distribution of the IPT and IPC weighted estimator and explicitly state the necessary terms to calculate a consistent estimator of this variance. We compare the performance of the robust and consistent variance estimators in an application based on routine health care data and in a simulation study. The simulation reveals no substantial differences between the 2 estimators in medium and large data sets with no unmeasured confounding, but the consistent variance estimator performs poorly in small samples or under unmeasured confounding, if the number of confounders is large. We thus conclude that the robust estimator is more appropriate for all practical purposes.

Comparison of multiple imputation and two-phase logistic regression to analyse two-phase case–control studies with rich phase 1: a simulation study

ABSTRACT Two-phase case–control studies cope with the problem of confounding by obtaining required additional information for a subset (phase 2) of all individuals (phase 1). Nowadays, studies with rich phase 1 data are available where only few unmeasured confounders need to be obtained in phase 2. The extended conditional maximum likelihood (ECML) approach in two-phase logistic regression is a novel method to analyse such data. Alternatively, two-phase case–control studies can be analysed by multiple imputation (MI), where phase 2 information for individuals included in phase 1 is treated as missing. We conducted a simulation of two-phase studies, where we compared the performance of ECML and MI in typical scenarios with rich phase 1. Regarding exposure effect, MI was less biased and more precise than ECML. Furthermore, ECML was sensitive against misspecification of the participation model. We therefore recommend MI to analyse two-phase case–control studies in situations with rich phase 1 data.