Nonproportional hazards and unobserved heterogeneity in clustered survival data: When can we tell the difference?

Multivariate survival data are frequently encountered in biomedical applications in the form of clustered failures (or recurrent events data). A popular way of analyzing such data is by using shared frailty models, which assume that the proportional hazards assumption holds conditional on an unobserved cluster-specific random effect. Such models are often incorporated in more complicated joint models in survival analysis. If the random effect distribution has finite expectation, then the conditional proportional hazards assumption does not carry over to the marginal models. It has been shown that, for univariate data, this makes it impossible to distinguish between the presence of unobserved heterogeneity (eg, due to missing covariates) and marginal nonproportional hazards. We show that time-dependent covariate effects may falsely appear as evidence in favor of a frailty model also in the case of clustered failures or recurrent events data, when the cluster size or number of recurrent events is small. When true unobserved heterogeneity is present, the presence of nonproportional hazards leads to overestimating the frailty effect. We show that this phenomenon is somewhat mitigated as the cluster size grows. We carry out a simulation study to assess the behavior of test statistics and estimators for frailty models in such contexts. The gamma, inverse Gaussian, and positive stable shared frailty models are contrasted using a novel software implementation for estimating semiparametric shared frailty models. Two main questions are addressed in the contexts of clustered failures and recurrent events: whether covariates with a time-dependent effect may appear as indication of unobserved heterogeneity and whether the additional presence of unobserved heterogeneity can be detected in this case. Finally, the practical implications are illustrated in a real-world data analysis example.     

Routine diagnostics for neural antibodies,clinical correlates,treatment and functional outcome

Objective

To determine frequencies, interlaboratory reproducibility, clinical ratings, and prognostic implications of neural antibodies in a routine laboratory setting in patients with suspected neuropsychiatric autoimmune conditions.

Methods

Earliest available samples from 10,919 patients were tested for a broad panel of neural antibodies. Sera that reacted with leucine-rich glioma-inactivated protein 1 (LGI1), contactin-associated protein-2 (CASPR2), or the voltage-gated potassium channel (VGKC) complex were retested for LGI1 and CASPR2 antibodies by another laboratory. Physicians in charge of patients with positive antibody results retrospectively reported on clinical, treatment, and outcome parameters.

Results

Positive results were obtained for 576 patients (5.3%). Median disease duration was 6 months (interquartile range 0.6–46 months). In most patients, antibodies were detected both in CSF and serum. However, in 16 (28%) patients with N-methyl-d-aspartate receptor (NMDAR) antibodies, this diagnosis could be made only in cerebrospinal fluid (CSF). The two laboratories agreed largely on LGI1 and CASPR2 antibody diagnoses (κ = 0.95). The clinicians (413 responses, 71.7%) rated two-thirds of the antibody-positive patients as autoimmune. Antibodies against the α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor (AMPAR), NMDAR (CSF or high serum titer), γ-aminobutyric acid-B receptor (GABABR), and LGI1 had ≥ 90% positive ratings, whereas antibodies against the glycine receptor, VGKC complex, or otherwise unspecified neuropil had ≤ 40% positive ratings. Of the patients with surface antibodies, 64% improved after ≥ 3 months, mostly with ≥ 1 immunotherapy intervention.

Conclusions

This novel approach starting from routine diagnostics in a dedicated laboratory provides reliable and useful results with therapeutic implications. Counseling should consider clinical presentation, demographic features, and antibody titers of the individual patient.

     

In vitro study of the sealing ability of four retrograde filling materials

Abstract This in vitro study utilized India ink dye after clearing to evaluate the extent of apical microleakage following reverse filling procedures. Forty single-rooted human teeth divided into 4 groups were chemomechanically prepared and obturated, using the lateral condensation technique with gutta-percha and Grossman sealer. Following obturation, an apicoectomy was performed and retrograde cavities were filled with 4 different materials: Group A: amalgam and varnish; Group B: EBA cement; Group C: Ketac-cem®; Group D: hot-burnished gutta-percha. All teeth were immersed in India ink, decalcified, cleaned, examined through a stereomicroscope, and the depth of linear dye penetration was measured. Statistical analysis showed significantly less dye penetration with EBA cement and amalgam with varnish than with Ketac-cem and hot-burnished gutta-percha.     

Benfotiamine prevents macro- and microvascular endothelial dysfunction and oxidative stress following a meal rich in advanced glycation end products in individuals with type 2 diabetes

OBJECTIVE: Diabetes is characterized by marked postprandial endothelial dysfunction induced by hyperglycemia, hypertriglyceridemia, advanced glycation end products (AGEs), and dicarbonyls (e.g., methylglyoxal [MG]). In vitro hyperglycemia-induced MG formation and endothelial dysfunction could be blocked by benfotiamine, but in vivo effects of benfotiamine on postprandial endothelial dysfunction and MG synthesis have not been investigated in humans until now. RESEARCH DESIGN AND METHODS: Thirteen people with type 2 diabetes were given a heat-processed test meal with a high AGE content (HAGE; 15.100 AGE kU, 580 kcal, 54 g protein, 17 g lipids, and 48 g carbohydrates) before and after a 3-day therapy with benfotiamine (1,050 mg/day). Macrovascular flow-mediated dilatation (FMD) and microvascular reactive hyperemia, along with serum markers of endothelial disfunction (E-selectin, vascular cell adhesion molecule-1, and intracellular adhesion molecule-1), oxidative stress, AGE, and MG were measured during both test meal days after an overnight fast and then at 2, 4, and 6 h postprandially. RESULTS: The HAGE induced a maximum reactive hyperemia decrease of -60.0% after 2 h and a maximum FMD impairment of -35.1% after 4 h, without affecting endothelium-independent vasodilatation. The effects of HAGE on both FMD and reactive hyperemia were completely prevented by benfotiamine. Serum markers of endothelial dysfunction and oxidative stress, as well as AGE, increased after HAGE. These effects were significantly reduced by benfotiamine. CONCLUSIONS: Our study confirms micro- and macrovascular endothelial dysfunction accompanied by increased oxidative stress following a real-life, heat-processed, AGE-rich meal in individuals with type 2 diabetes and suggests benfotiamine as a potential treatment.     

A supra-annular malposition of the Perceval S sutureless aortic valve: the 'χ-movement' removal technique and subsequent reimplantation

The Perceval S sutureless valve prosthesis has recently been introduced as a new biological aortic valve prosthesis, but a specific learning curve is required, as for every cardiac surgical centre dealing with a new technique. After the removal of the stenotic valve, the prosthetic valve is correctly positioned within the mildly decalcified aortic annulus. When a supra-annular malposition occurs, due to an excessively rapid release of the prosthesis in the aorta or incomplete annular visualization, the Perceval S valve can safely be removed even after balloon dilation. The procedure performed is a 'χ-movement' with the aid of anatomical forceps. If the prosthesis does not show any malformation after the procedure, it can be reimplanted in the correct intra-annular position.     

Allogenous vein graft as vascular access for hemodialysis - lost battle?

Purpose: The purpose of this paper is to assess a long-term outcome of allogenous vein grafts (ALVG) as vascular access for hemodialysis. Materials and methods: For nearly eight years (between 9/2002 and 9/2011) a total of 78 patients with 112 ALVGs were involved in the study. The register included 46 women and 32 men, mean age 66.1 ± 11.2 years; range 20-88 years. The patient database was retrospectively reviewed and statistical processing was performed. Results: Almost all ALVGs were treated by PTA or surgically, very often repeatedly. The number of radiologic interventions was 316, the number of surgical procedures 31. Mean follow-up time was 795 days, range 28-3522 days. Thirty-five patients died of unrelated causes, nineteen with functional graft, fourteen patients were lost to follow-up. Forty ALVGs failed for various reasons, mostly because of occlusion. Only one patient underwent successful renal transplantation, no patient converted to peritoneal dialysis. Thirty-seven ALVGs remain correctly functioning. Primary patency rates at 6, 12, and 24 months were 81 ± 5%, 63 ± 5%, and 34 ± 2% respectively. Secondary patency rates at 6, 12, and 24 months were 96 ± 2%, 82 ± 4%, and 65 ± 5% respectively. Conclusions: Allogenous vein grafts, in spite of the high number of necessary radiologic and surgical interventions and reinterventions, show acceptable clinical usability and durability, comparable with other types of prosthetic grafts.     

Submillisecond unmasked subliminal visual stimuli evoke electrical brain responses

Subliminal perception is strongly associated to the processing of meaningful or emotional information and has mostly been studied using visual masking. In this study, we used high density 256‐channel EEG coupled with an liquid crystal display (LCD) tachistoscope to characterize the spatio‐temporal dynamics of the brain response to visual checkerboard stimuli (Experiment 1) or blank stimuli (Experiment 2) presented without a mask for 1 ms (visible), 500 µs (partially visible), and 250 µs (subliminal) by applying time‐wise, assumption‐free nonparametric randomization statistics on the strength and on the topography of high‐density scalp‐recorded electric field. Stimulus visibility was assessed in a third separate behavioral experiment. Results revealed that unmasked checkerboards presented subliminally for 250 µs evoked weak but detectable visual evoked potential (VEP) responses. When the checkerboards were replaced by blank stimuli, there was no evidence for the presence of an evoked response anymore. Furthermore, the checkerboard VEPs were modulated topographically between 243 and 296 ms post‐stimulus onset as a function of stimulus duration, indicative of the engagement of distinct configuration of active brain networks. A distributed electrical source analysis localized this modulation within the right superior parietal lobule near the precuneus. These results show the presence of a brain response to submillisecond unmasked subliminal visual stimuli independently of their emotional saliency or meaningfulness and opens an avenue for new investigations of subliminal stimulation without using visual masking. Hum Brain Mapp 36:1470–1483, 2015. © 2014 Wiley Periodicals, Inc.