Recent data indicated that reduction of the aminoglycoside dose might yield selective high-frequency hearing loss. Such a model is relevant for studies related to hybrid cochlear
implant devices, for example, with respect to preservation Wnt pathway of residual hearing.
Methods: Guinea pigs received an electrode for chronic recording of compound action potentials to tones to assess thresholds. They were treated with a coadministration of kanamycin (200 mg/kg) and furosemide (100 mg/kg), after which, the animals were sacrificed for histologic analysis at 2, 4, or 7 weeks.
Results: After 2 to 7 weeks threshold shifts were greater than 50 dB for 8 to 16 kHz in 15 of 17 animals, whereas threshold shifts at 2 kHz or lower were less than 50 dB in 13 animals. Major threshold shifts occurred the first 2 to 4 days; subsequently, some spontaneous recovery occurred and, after 2-3 weeks thresholds, remained stable. Inner hair cell loss still progressed between 2 this website and 4 weeks in the most basal cochlear region; thereafter, hair cell loss was stable.
Conclusion: An appropriate animal model for selective severe high-frequency hearing loss was obtained, which is stable at 4 weeks after ototoxic treatment.”
“Background: For the analysis of length-of-stay (LOS) data, which is characteristically
right-skewed, a number of statistical estimators have been proposed as alternatives to the traditional ordinary least squares (OLS) regression with log dependent variable.
Methods: Using a cohort of patients identified in the Australian and New Zealand Intensive Care Society Adult Patient www.selleckchem.com/products/xmu-mp-1.html Database, 2008-2009, 12 different methods were used for estimation of intensive care (ICU) length of stay. These encompassed risk-adjusted regression analysis of firstly: log LOS using OLS, linear mixed model [LMM], treatment effects, skew-normal and skew-t models; and secondly: unmodified (raw) LOS via OLS, generalised linear models [GLMs] with log-link and 4 different distributions [Poisson, gamma, negative binomial and inverse-Gaussian], extended estimating equations
[EEE] and a finite mixture model including a gamma distribution. A fixed covariate list and ICU-site clustering with robust variance were utilised for model fitting with split-sample determination (80%) and validation (20%) data sets, and model simulation was undertaken to establish over-fitting (Copas test). Indices of model specification using Bayesian information criterion [BIC: lower values preferred] and residual analysis as well as predictive performance (R-2, concordance correlation coefficient (CCC), mean absolute error [MAE]) were established for each estimator.
Results: The data-set consisted of 111663 patients from 131 ICUs; with mean(SD) age 60.6(18.8) years, 43.0% were female, 40.7% were mechanically ventilated and ICU mortality was 7.8%. ICU length-of-stay was 3.4(5.1) (median 1.8, range (0.