A parametric additive hazard model for time-to-event analysis.

BACKGROUND: In recent years, the use of non- and semi-parametric models which estimate hazard ratios for analysing time-to-event outcomes is continuously criticized in terms of interpretation, technical implementation, and flexibility. Hazard ratios in particular are critically discussed for their misleading interpretation as relative risks and their non-collapsibility. Additive hazard models do not have these drawbacks but are rarely used because they assume a non- or semi-parametric additive hazard which renders computation and interpretation complicated. METHODS: As a remedy, we propose a new parametric additive hazard model that allows results to be reported on the original time rather than on the hazard scale. Being an essentially parametric model, survival, hazard and probability density functions are directly available. Parameter estimation is straightforward by maximizing the log-likelihood function. RESULTS: Applying the model to different parametric distributions in a simulation study and in an exemplary application using data from a study investigating medical care to lung cancer patients, we show that the approach works well in practice. CONCLUSIONS: Our proposed parametric additive hazard model can serve as a powerful tool to analyze time-to-event outcomes due to its simple interpretation, flexibility and facilitated parameter estimation.

Trajectories of Body Mass Index from Childhood to Young Adulthood among Patients with Type 1 Diabetes-A Longitudinal Group-Based Modeling Approach Based on the DPV Registry.

OBJECTIVE: To identify distinct longitudinal patterns of body mass index (BMI) z score in type 1 diabetes from childhood to young adulthood and secondly to determine sex differences as well as associated clinical covariates. STUDY DESIGN: A total of 5665 patients with type 1 diabetes (51% male) with follow-up from 8 to 20 years of age from the multicenter diabetes prospective registry DPV were studied (baseline diabetes duration ≥1 years, BMI z score aggregated per year of life). Latent class growth modeling (SAS: PROC TRAJ) was applied to analyze BMI z score over time. RESULTS: Six distinct BMI z score trajectories were identified (group 1: 7% of patients, group 2: 22%, group 3: 20%, group 4: 16%, group 5: 25%, and group 6: 10%). Group 1, 2, 5, and 6 had an almost stable BMI z score, either in the low, near-normal, high stable, or chronic overweight range. Group 3 (60% girls) increased their BMI during puberty, whereas group 4 (65% boys) had a BMI decrease. Similar patterns were observed for girls only, whereas boys followed nearly stable trajectories without fluctuation over time. Between the near-normal and the other groups, significant differences (P < .05) in sex ratio, migration background, mental health, height z score, glycated hemoglobin A1c, diabetes treatment, dyslipidemia, hypertension, and smoking were observed. CONCLUSIONS: In youth with type 1 diabetes, a great heterogeneity of BMI z score trajectories exists that highlight the importance of personalized sex-specific intervention programs for subjects at risk for unfavorable BMI development.

A description of clinician reported diagnosis of type 2 diabetes and other non-type 1 diabetes included in a large international multicentered pediatric diabetes registry (SWEET).

BACKGROUND: Although type 1 diabetes (T1D) remains the most frequent form of diabetes in individuals aged less than 20 years at onset, other forms of diabetes are being increasingly recognized. OBJECTIVES: To describe the population of children with other forms of diabetes (non-type 1) included in the multinational SWEET (Better control in Pediatric and Adolescent diabeteS: Working to crEate CEnTers of Reference) database for children with diabetes. METHODS: Cases entered in the SWEET database are identified by their physician as T1D, type 2 diabetes (T2D) and other types of diabetes according to the ISPAD classification. Etiologic subgroups are provided for other types of diabetes. Descriptive analyses were tabulated for age at onset, gender, daily insulin doses, and hemoglobin A1c (A1C) for each type and subtype of diabetes and when possible, values were compared. RESULTS: Of the 27 104 patients included in this report, 95.5% have T1D, 1.3% T2D, and 3.2% other forms of diabetes. The two most frequent etiologies for other forms of diabetes were maturity onset diabetes of the young (MODY) (n = 351) and cystic fibrosis-related diabetes (CFRD) (n = 193). The cause was unknown or unreported in 10% of other forms of diabetes. Compared with T1D, children with T2D and CFRD were diagnosed at an older age, took less insulin and had lower A1C (all P < .0001). CONCLUSION: In centers included in SWEET, forms of diabetes other than type 1 remain rare and at times difficult to characterize. Sharing clinical information and outcome between SWEET centers on those rare forms of diabetes has the potential to improve management and outcome.

Genetic and environmental risk factors for submucous cleft palate.

A multifactorial aetiology with genetic and environmental factors is assumed for orofacial clefts. Submucous cleft palate (SMCP), a subgroup of cleft palates with insufficient median fusion of the muscles of the soft palate hidden under the mucosa, has a prevalence of 1:1,250-1:5,000. We described the prevalence of risk factors among 103 German patients with the subtype SMCP and genotyped 24 single nucleotide polymorphisms (SNPs) from 12 candidate genes for orofacial clefts. Analysis of risk factors yielded a positive history for maternal cigarette smoking during pregnancy in 25.2% of the patients, and this was significantly more frequent than in the normal population. The group of patients differed in allele frequencies at SNP rs3917192 of the gene TGFB3 (nominal P = 0.053) and at SNP rs5752638 of the gene MN1 (nominal P = 0.075) compared with 279 control individuals. Our results indicate a potential role of maternal smoking during pregnancy for the formation of SMCP. The analysis of genetic variants hints at the contribution of TGFB3 and MN1 in the aetiology of SMCPs.