Stochastic search for semiparametric linear regression models

This paper introduces and analyzes a stochastic search method for parameter estimation in linear regression models in the spirit of Beran and Millar [Ann. Statist. 15(3) (1987) 1131–1154]. The idea is to generate a random finite subset of a parameter space which will automatically contain points which are very close to an unknown true parameter. The motivation for this procedure comes from recent work of Dümbgen et al. [Ann. Statist. 39(2) (2011) 702–730] on regression models with log-concave error distributions.

Bacillus Calmette-Guérin Failure in Patients with Non-Muscle-invasive Urothelial Carcinoma of the Bladder May Be Due to the Urologist's Failure to Detect Urothelial Carcinoma of the Upper Urinary Tract and Urethra

BACKGROUND: Various reasons exist for so-called bacillus Calmette-Guérin (BCG) failure in patients with non-muscle-invasive urothelial bladder carcinoma (NMIBC). OBJECTIVE: To explore whether urothelial carcinoma of the upper urinary tract (UUT) and/or prostatic urethra may be a cause for BCG failure. DESIGN, SETTING, AND PARTICIPANTS: Retrospective analysis of 110 patients with high-risk NMIBC repeatedly treated with intravesical BCG, diagnosed with disease recurrence, and followed for a median time of 9.1 yr. INTERVENTION: Two or more intravesical BCG induction courses without maintenance. OUTCOME MEASUREMENTS AND STATISTICAL ANALYSIS: Primary outcome was pattern of disease recurrence (BCG failure) within the urinary tract categorised into UUT and/or urethral carcinoma (with or without intravesical recurrence), and intravesical recurrence alone. Secondary outcome was survival. Predictors of UUT and/or urethral carcinoma and the effect of pattern of disease recurrence on cancer-specific survival were assessed with multivariable Cox regression analysis adjusting for multiple clinical and tumour characteristics. RESULTS AND LIMITATIONS: Of the 110 patients, 57 (52%) had UUT and/or urethral carcinoma (with or without intravesical recurrence), and 53 (48%) had intravesical recurrence alone. In patients with UUT and/or urethral carcinoma, bladder carcinoma in situ (Tis) before the first and second BCG course was present in 42 of 57 (74%) and 47 of 57 (82%) patients, respectively. On multivariable analysis, bladder Tis before the first and/or second BCG course was the only independent predictor of UUT and/or urethral carcinoma. Of the 110 patients, 69 (63%) were alive at last follow-up visit, 18 (16%) had died due to metastatic urothelial carcinoma, and 23 (21%) had died of other causes. Pattern of disease recurrence within the urinary tract was not an independent predictor of cancer-specific survival. Main study limitations were retrospective design and limited power for survival analysis. CONCLUSIONS: In our patients with high-risk NMIBC failing after two or more courses of intravesical BCG, UUT and/or urethral carcinoma was detected in >50% of the cases during follow-up. The vast majority of these patients had bladder Tis before the first and/or second BCG course. In patients experiencing the so-called BCG failure, a diagnostic work-up of UUT and prostatic urethra should always be performed to exclude urothelial carcinoma before additional intravesical therapy or even a radical cystectomy is considered.

Fluid intake and all-cause mortality, cardiovascular mortality, and kidney function: a population-based longitudinal cohort study

BACKGROUND Drinking eight glasses of fluid or water each day is widely believed to improve health, but evidence is sparse and conflicting. We aimed to investigate the association between fluid consumption and long-term mortality and kidney function. METHODS We conducted a longitudinal analysis within a prospective, population-based cohort study of 3858 men and women aged 49 years or older residing in Australia. Daily fluid intake from food and beverages not including water was measured using a food frequency questionnaire. We did multivariable adjusted Cox proportional hazard models for all-cause and cardiovascular mortality and a boot-strapping procedure for estimated glomerular filtration rate (eGFR). RESULTS Upper and lower quartiles of daily fluid intake corresponded to >3 L and <2 L, respectively. During a median follow-up of 13.1 years (total 43 093 years at risk), 1127 deaths (26.1 per 1000 years at risk) including 580 cardiovascular deaths (13.5 per 1000 years at risk) occurred. Daily fluid intake (per 250 mL increase) was not associated with all-cause [adjusted hazard ratio (HR) 0.99 (95% CI 0.98-1.01)] or cardiovascular mortality [HR 0.98 (95% CI 0.95-1.01)]. Overall, eGFR reduced by 2.2 mL/min per 1.73 m(2) (SD 10.9) in the 1207 (31%) participants who had repeat creatinine measurements and this was not associated with fluid intake [adjusted regression coefficient 0.06 mL/min/1.73 m(2) per 250 mL increase (95% CI -0.03 to 0.14)]. CONCLUSIONS Fluid intake from food and beverages excluding water is not associated with improved kidney function or reduced mortality.

Adherence as a Predictor of the Development of Class-Specific Resistance Mutations: The Swiss HIV Cohort Study

Background Non-adherence is one of the strongest predictors of therapeutic failure in HIV-positive patients. Virologic failure with subsequent emergence of resistance reduces future treatment options and long-term clinical success. Methods Prospective observational cohort study including patients starting new class of antiretroviral therapy (ART) between 2003 and 2010. Participants were naïve to ART class and completed ≥1 adherence questionnaire prior to resistance testing. Outcomes were development of any IAS-USA, class-specific, or M184V mutations. Associations between adherence and resistance were estimated using logistic regression models stratified by ART class. Results Of 314 included individuals, 162 started NNRTI and 152 a PI/r regimen. Adherence was similar between groups with 85% reporting adherence ≥95%. Number of new mutations increased with increasing non-adherence. In NNRTI group, multivariable models indicated a significant linear association in odds of developing IAS-USA (odds ratio (OR) 1.66, 95% confidence interval (CI): 1.04-2.67) or class-specific (OR 1.65, 95% CI: 1.00-2.70) mutations. Levels of drug resistance were considerably lower in PI/r group and adherence was only significantly associated with M184V mutations (OR 8.38, 95% CI: 1.26-55.70). Adherence was significantly associated with HIV RNA in PI/r but not NNRTI regimens. Conclusion Therapies containing PI/r appear more forgiving to incomplete adherence compared with NNRTI regimens, which allow higher levels of resistance, even with adherence above 95%. However, in failing PI/r regimens good adherence may prevent accumulation of further resistance mutations and therefore help to preserve future drug options. In contrast, adherence levels have little impact on NNRTI treatments once the first mutations have emerged.

Non-AIDS defining cancers in the D:A:D Study - time trends and predictors of survival: a cohort study

Background Non-AIDS defining cancers (NADC) are an important cause of morbidity and mortality in HIV-positive individuals. Using data from a large international cohort of HIV-positive individuals, we described the incidence of NADC from 2004–2010, and described subsequent mortality and predictors of these. Methods Individuals were followed from 1st January 2004/enrolment in study, until the earliest of a new NADC, 1st February 2010, death or six months after the patient’s last visit. Incidence rates were estimated for each year of follow-up, overall and stratified by gender, age and mode of HIV acquisition. Cumulative risk of mortality following NADC diagnosis was summarised using Kaplan-Meier methods, with follow-up for these analyses from the date of NADC diagnosis until the patient’s death, 1st February 2010 or 6 months after the patient’s last visit. Factors associated with mortality following NADC diagnosis were identified using multivariable Cox proportional hazards regression. Results Over 176,775 person-years (PY), 880 (2.1%) patients developed a new NADC (incidence: 4.98/1000PY [95% confidence interval 4.65, 5.31]). Over a third of these patients (327, 37.2%) had died by 1st February 2010. Time trends for lung cancer, anal cancer and Hodgkin’s lymphoma were broadly consistent. Kaplan-Meier cumulative mortality estimates at 1, 3 and 5 years after NADC diagnosis were 28.2% [95% CI 25.1-31.2], 42.0% [38.2-45.8] and 47.3% [42.4-52.2], respectively. Significant predictors of poorer survival after diagnosis of NADC were lung cancer (compared to other cancer types), male gender, non-white ethnicity, and smoking status. Later year of diagnosis and higher CD4 count at NADC diagnosis were associated with improved survival. The incidence of NADC remained stable over the period 2004–2010 in this large observational cohort. Conclusions The prognosis after diagnosis of NADC, in particular lung cancer and disseminated cancer, is poor but has improved somewhat over time. Modifiable risk factors, such as smoking and low CD4 counts, were associated with mortality following a diagnosis of NADC.

Impact of Switching From Zidovudine to Tenofovir Disoproxil Fumarate on Bone Mineral Density and Markers of Bone Metabolism in Virologically Suppressed HIV-1 Infected Patients; A Substudy of the PREPARE Study

Context: In virologically suppressed, antiretroviral-treated patients, the effect of switching to tenofovir (TDF) on bone biomarkers compared to patients remaining on stable antiretroviral therapy is unknown. Methods: We examined bone biomarkers (osteocalcin [OC], procollagen type 1 amino-terminal propeptide, and C-terminal cross-linking telopeptide of type 1 collagen) and bone mineral density (BMD) over 48 weeks in virologically suppressed patients (HIV RNA < 50 copies/ml) randomized to switch to TDF/emtricitabine (FTC) or remain on first-line zidovudine (AZT)/lamivudine (3TC). PTH was also measured. Between-group differences in bone biomarkers and associations between change in bone biomarkers and BMD measures were assessed by Student's t tests, Pearson correlation, and multivariable linear regression, respectively. All data are expressed as mean (SD), unless otherwise specified. Results: Of 53 subjects (aged 46.0 y; 84.9% male; 75.5% Caucasian), 29 switched to TDF/FTC. There were reductions in total hip and lumbar spine BMD in those switching to TDF/FTC (total hip, TDF/FTC, −1.73 (2.76)% vs AZT/3TC, −0.39 (2.41)%; between-group P = .07; lumbar spine, TDF/FTC, −1.50 (3.49)% vs AZT/3TC, +0.25 (2.82)%; between-group P = .06), but they did not reach statistical significance. Greater declines in lumbar spine BMD correlated with greater increases in OC (r = −0.28; P = .05). The effect of TDF/FTC on bone biomarkers remained significant when adjusted for baseline biomarker levels, gender, and ethnicity. There was no difference in change in PTH levels over 48 weeks between treatment groups (between-group P = .23). All biomarkers increased significantly from weeks 0 to 48 in the switch group, with no significant change in those remaining on AZT/3TC (between-group, all biomarkers, P < .0001). Conclusion: A switch to TDF/FTC compared to remaining on a stable regimen is associated with increases in bone turnover that correlate with reductions in BMD, suggesting that TDF exposure directly affects bone metabolism in vivo.

Potentially avoidable 30-day hospital readmissions in medical patients: derivation and validation of a prediction model

IMPORTANCE Because effective interventions to reduce hospital readmissions are often expensive to implement, a score to predict potentially avoidable readmissions may help target the patients most likely to benefit. OBJECTIVE To derive and internally validate a prediction model for potentially avoidable 30-day hospital readmissions in medical patients using administrative and clinical data readily available prior to discharge. DESIGN Retrospective cohort study. SETTING Academic medical center in Boston, Massachusetts. PARTICIPANTS All patient discharges from any medical services between July 1, 2009, and June 30, 2010. MAIN OUTCOME MEASURES Potentially avoidable 30-day readmissions to 3 hospitals of the Partners HealthCare network were identified using a validated computerized algorithm based on administrative data (SQLape). A simple score was developed using multivariable logistic regression, with two-thirds of the sample randomly selected as the derivation cohort and one-third as the validation cohort. RESULTS Among 10 731 eligible discharges, 2398 discharges (22.3%) were followed by a 30-day readmission, of which 879 (8.5% of all discharges) were identified as potentially avoidable. The prediction score identified 7 independent factors, referred to as the HOSPITAL score: h emoglobin at discharge, discharge from an o ncology service, s odium level at discharge, p rocedure during the index admission, i ndex t ype of admission, number of a dmissions during the last 12 months, and l ength of stay. In the validation set, 26.7% of the patients were classified as high risk, with an estimated potentially avoidable readmission risk of 18.0% (observed, 18.2%). The HOSPITAL score had fair discriminatory power (C statistic, 0.71) and had good calibration. CONCLUSIONS AND RELEVANCE This simple prediction model identifies before discharge the risk of potentially avoidable 30-day readmission in medical patients. This score has potential to easily identify patients who may need more intensive transitional care interventions.

Individual risk factors for Mycoplasma hyopneumoniae infections in suckling pigs at the age of weaning.

BACKGROUND In recent years, the occurrence and the relevance of Mycoplasma hyopneumoniae infections in suckling pigs has been examined in several studies. Whereas most of these studies were focused on sole prevalence estimation within different age groups, follow-up of infected piglets or assessment of pathological findings, none of the studies included a detailed analysis of individual and environmental risk factors. Therefore, the aim of the present study was to investigate the frequency of M. hyopneumoniae infections in suckling pigs of endemically infected herds and to identify individual risk factors potentially influencing the infection status of suckling pigs at the age of weaning. RESULTS The animal level prevalence of M. hyopneumoniae infections in suckling pigs examined in three conventional pig breeding herds was 3.6% (41/1127) at the time of weaning. A prevalence of 1.2% was found in the same pigs at the end of their nursery period. In a multivariable Poisson regression model it was found that incidence rate ratios (IRR) for suckling pigs are significantly lower than 1 when teeth grinding was conducted (IRR: 0.10). Moreover, high temperatures in the piglet nest during the first two weeks of life (occasionally >40°C) were associated with a decrease of the probability of an infection (IRR: 0.23-0.40). Contrary, the application of PCV2 vaccines to piglets was associated with an increased infection risk (IRR: 9.72). CONCLUSIONS Since single infected piglets are supposed to act as initiators for the transmission of this pathogen in nursery and fattening pigs, the elimination of the risk factors described in this study should help to reduce the incidence rate of M. hyopneumoniae infections and thereby might contribute to a reduced probability of high prevalences in older pigs.

Cotrimoxazole prophylaxis is associated with reduced risk of incident tuberculosis in participants in the Swiss HIV Cohort Study

Cotrimoxazole reduces mortality in HIV-infected adults with tuberculosis (TB), and in vitro data suggest potential anti-mycobacterial activity of cotrimoxazole. We aimed to evaluate whether prophylaxis with cotrimoxazole is associated with a decreased risk of incident TB in SHCS participants. We determined the incidence of TB per 1000 person-years from January 1992 to December 2012. Rates were analyzed separately in participants with current or no previous antiretroviral treatment (ART) using Poisson regression adjusted for CD4 cell count, sex, region of origin, injecting drug use, and age. 13,431 cohort participants contributed 107,549 person-years follow-up; 182 patients had incident TB; 132 (73%) before and 50 (27%) after ART initiation. The multivariable incidence rate ratios for cumulative cotrimoxazole exposure per year for persons with no previous and current ART were 0.70 (95% CI 0.55-0.89) and 0.87 (0.74-1.0) respectively. Cotrimoxazole may prevent the development of TB among HIV-positive persons, especially among those with no previous ART.

Questionnaire-based study to assess the association between management practices and mastitis within tie-stall and free-stall dairy housing systems in Switzerland.

BACKGROUND Prophylactic measures are key components of dairy herd mastitis control programs, but some are only relevant in specific housing systems. To assess the association between management practices and mastitis incidence, data collected in 2011 by a survey among 979 randomly selected Swiss dairy farms, and information from the regular test day recordings from 680 of these farms was analyzed. RESULTS The median incidence of farmer-reported clinical mastitis (ICM) was 11.6 (mean 14.7) cases per 100 cows per year. The median annual proportion of milk samples with a composite somatic cell count (PSCC) above 200,000 cells/ml was 16.1 (mean 17.3) %. A multivariable negative binomial regression model was fitted for each of the mastitis indicators for farms with tie-stall and free-stall housing systems separately to study the effect of other (than housing system) management practices on the ICM and PSCC events (above 200,000 cells/ml). The results differed substantially by housing system and outcome. In tie-stall systems, clinical mastitis incidence was mainly affected by region (mountainous production zone; incidence rate ratio (IRR) = 0.73), the dairy herd replacement system (1.27) and farmers age (0.81). The proportion of high SCC was mainly associated with dry cow udder controls (IRR = 0.67), clean bedding material at calving (IRR = 1.72), using total merit values to select bulls (IRR = 1.57) and body condition scoring (IRR = 0.74). In free-stall systems, the IRR for clinical mastitis was mainly associated with stall climate/temperature (IRR = 1.65), comfort mats as resting surface (IRR = 0.75) and when no feed analysis was carried out (IRR = 1.18). The proportion of high SSC was only associated with hand and arm cleaning after calving (IRR = 0.81) and beef producing value to select bulls (IRR = 0.66). CONCLUSIONS There were substantial differences in identified risk factors in the four models. Some of the factors were in agreement with the reported literature while others were not. This highlights the multifactorial nature of the disease and the differences in the risks for both mastitis manifestations. Attempting to understand these multifactorial associations for mastitis within larger management groups continues to play an important role in mastitis control programs.

Tumor Regression Grading of Gastrointestinal Carcinomas after Neoadjuvant Treatment

Multimodal therapy concepts have been successfully implemented in the treatment of locally advanced gastrointestinal malignancies. The effects of neoadjuvant chemo- or radiochemotherapy such as scarry fibrosis or resorptive changes and inflammation can be determined by histopathological investigation of the subsequent resection specimen. Tumor regression grading (TRG) systems which aim to categorize the amount of regressive changes after cytotoxic treatment mostly refer onto the amount of therapy induced fibrosis in relation to residual tumor or the estimated percentage of residual tumor in relation to the previous tumor site. Commonly used TRGs for upper gastrointestinal carcinomas are the Mandard grading and the Becker grading system, e.g., and for rectal cancer the Dworak or the Rödel grading system, or other systems which follow similar definitions. Namely for gastro-esophageal carcinomas these TRGs provide important prognostic information since complete or subtotal tumor regression has shown to be associated with better patient's outcome. The prognostic value of TRG may even exceed those of currently used staging systems (e.g., TNM staging) for tumors treated by neoadjuvant therapy. There have been some limitations described regarding interobserver variability especially in borderline cases, which may be improved by standardization of work up of resection specimen and better training of histopathologic determination of regressive changes. It is highly recommended that TRG should be implemented in every histopathological report of neoadjuvant treated gastrointestinal carcinomas. The aim of this review is to disclose the relevance of histomorphological TRG to accomplish an optimal therapy for patients with gastrointestinal carcinomas.

Tumor regression grade of urothelial bladder cancer after neoadjuvant chemotherapy: a novel and successful strategy to predict survival.

Histopathologic tumor regression grades (TRGs) after neoadjuvant chemotherapy predict survival in different cancers. In bladder cancer, corresponding studies have not been conducted. Fifty-six patients with advanced invasive urothelial bladder cancer received neoadjuvant chemotherapy before cystectomy and lymphadenectomy. TRGs were defined as follows: TRG1: complete tumor regression; TRG2: >50% tumor regression; TRG3: 50% or less tumor regression. Separate TRGs were assigned for primary tumors and corresponding lymph nodes. The prognostic impact of these 2 TRGs, the highest (dominant) TRG per patient, and competing tumor features reflecting tumor regression (ypT/ypN stage, maximum diameter of the residual tumor) were determined. Tumor characteristics in initial transurethral resection of the bladder specimens were tested for response prediction. The frequency of TRGs 1, 2, and 3 in the primary tumors were n=16, n=19, and n=21; corresponding data from the lymph nodes were n=31, n=9, and n=16. Interobserver agreement in determination of the TRG was strong (κ=0.8). Univariately, all evaluated parameters were significantly (P≤0.001) related to overall survival; however, the segregation of the Kaplan-Meier curves was best for the dominant TRG. In multivariate analysis, only dominant TRG predicted overall survival independently (P=0.035). In transurethral resection specimens of the chemotherapy-naive bladder cancer, the only tumor feature with significant (P<0.03) predictive value for therapy response was a high proliferation rate. In conclusion, among all parameters reflecting tumor regression, the dominant TRG was the only independent risk factor. A favorable chemotherapy response is associated with a high proliferation rate in the initial chemotherapy-naive bladder cancer. This feature might help personalize neoadjuvant chemotherapy.

Comparison of partial least squares regression and principal component regression for pelvic shape prediction.

This paper studied two different regression techniques for pelvic shape prediction, i.e., the partial least square regression (PLSR) and the principal component regression (PCR). Three different predictors such as surface landmarks, morphological parameters, or surface models of neighboring structures were used in a cross-validation study to predict the pelvic shape. Results obtained from applying these two different regression techniques were compared to the population mean model. In almost all the prediction experiments, both regression techniques unanimously generated better results than the population mean model, while the difference on prediction accuracy between these two regression methods is not statistically significant (α=0.01).

3D volumetric intensity reconsturction from 2D x-ray images using partial least squares regression

Reconstruction of shape and intensity from 2D x-ray images has drawn more and more attentions. Previously introduced work suffers from the long computing time due to its iterative optimization characteristics and the requirement of generating digitally reconstructed radiographs within each iteration. In this paper, we propose a novel method which uses a patient-specific 3D surface model reconstructed from 2D x-ray images as a surrogate to get a patient-specific volumetric intensity reconstruction via partial least squares regression. No DRR generation is needed. The method was validated on 20 cadaveric proximal femurs by performing a leave-one-out study. Qualitative and quantitative results demonstrated the efficacy of the present method. Compared to the existing work, the present method has the advantage of much shorter computing time and can be applied to both DXA images as well as conventional x-ray images, which may hold the potentials to be applied to clinical routine task such as total hip arthroplasty (THA).