Cross-validation approaches for penalized Cox regression.

Cross-validation is the most common way of selecting tuning parameters in penalized regression, but its use in penalized Cox regression models has received relatively little attention in the literature. Due to its partial likelihood construction, carrying out cross-validation for Cox models is not straightforward, and there are several potential approaches for implementation. Here, we propose a new approach based on cross-validating the linear predictors of the Cox model and compare it to approaches that have been proposed elsewhere. We show that the proposed approach offers an attractive balance of performance and numerical stability, and illustrate these advantages using simulated data as well as analyzing a high-dimensional study of gene expression and survival in lung cancer patients.

5-Flucytosine Longitudinal Antifungal Susceptibility Testing of Cryptococcus neoformans: A Substudy of the EnACT Trial Testing Oral Amphotericin.

Background: The EnACT trial was a phase 2 randomized clinical trial conducted in Uganda, which evaluated a novel orally delivered lipid nanocrystal (LNC) amphotericin B in combination with flucytosine for the treatment of cryptococcal meningitis. When flucytosine (5FC) is used as monotherapy in cryptococcosis, 5FC can induce resistant Cryptococcus mutants. Oral amphotericin B uses a novel drug delivery mechanism, and we assessed whether resistance to 5FC develops during oral LNC-amphotericin B therapy. Methods: We enrolled Ugandans with HIV diagnosed with cryptococcal meningitis and who were randomized to receive 5FC and either standard intravenous (IV) amphotericin B or oral LNC-amphotericin B. We used broth microdilution to measure the minimum inhibitory concentration (MIC) of the first and last cryptococcal isolates in each participant. Breakpoints are inferred from 5FC in Candida albicans. We measured cerebral spinal fluid (CSF) 5FC concentrations by liquid chromatography and tandem mass spectrometry. Results: Cryptococcus 5FC MIC50 was 4 µg/mL, and MIC90 was 8 µg/mL. After 2 weeks of therapy, there was no evidence of 5FC resistance developing, defined as a >4-fold change in susceptibility in any Cryptococcus isolate tested. The median CSF 5FC concentration to MIC ratio (interquartile range) was 3.0 (1.7-5.5) µg/mL. There was no association between 5FC/MIC ratio and early fungicidal activity of the quantitative rate of CSF yeast clearance (R2 = 0.004; P = .63). Conclusions: There is no evidence of baseline resistance to 5FC or incident resistance during combination therapy with oral or IV amphotericin B in Uganda. Oral amphotericin B can safely be used in combination with 5FC.

Achieving Normoglycemia With Tirzepatide: Analysis of SURPASS 1-4 Trials.

OBJECTIVE: Tirzepatide is a novel single-molecule glucose-dependent insulinotropic polypeptide/glucagon-like peptide 1 receptor agonist, which demonstrated unprecedented improvements in glycemic control and body weight reduction, in the SURPASS phase 3 program. In this exploratory analysis, we aimed to characterize tirzepatide-treated participants who achieved HbA1c <5.7% and evaluate changes in clinical markers associated with long-term cardiometabolic health. RESEARCH DESIGN AND METHODS: Baseline characteristics and change from baseline to week 40 for several efficacy and safety parameters were analyzed according to HbA1c attainment category (<5.7%, 5.7-6.5%, and >6.5%) using descriptive statistics in participants taking ≥75% of treatment doses, without rescue medication, in the SURPASS 1-4 trials (N = 3,229). Logistic regression models with tirzepatide doses adjusted as a covariate were used to obtain odds ratios and assess the impact of patient characteristics achieving an HbA1c <5.7%. RESULTS: Tirzepatide-treated participants who achieved HbA1c <5.7% were slightly younger, with a shorter duration of diabetes and lower HbA1c value at baseline compared with those who did not achieve HbA1c <5.7%. In addition, they showed greater improvements in HbA1c, body weight, waist circumference, blood pressure, liver enzymes, and lipid parameters without increasing hypoglycemia risk. CONCLUSIONS: Normoglycemia was unprecedently achieved in a significant proportion of participants in the SURPASS clinical program, without increasing hypoglycemia risk, and was associated with an overall improvement in metabolic health.

Oral Lipid Nanocrystal Amphotericin B for Cryptococcal Meningitis: A Randomized Clinical Trial.

BACKGROUND: Amphotericin B is the gold standard treatment for severe mycoses. A new orally delivered, less-toxic formulation of amphotericin has been developed. METHODS: In our randomized clinical trial, we tested oral lipid nanocrystal (LNC) amphotericin B (MAT2203, Matinas Biopharma) vs intravenous (IV) amphotericin for human immunodeficiency virus-associated cryptococcal meningitis in 4 sequential cohorts. Two pilot cohorts assessed safety and tolerability (n = 10 each), and 2 cohorts assessed efficacy with/without 2 IV loading doses (n = 40 each). The experimental arm received 1.8 g/d oral LNC amphotericin through 2 weeks with 100 mg/kg/d flucytosine, then 1.2 g/d LNC amphotericin through 6 weeks. The randomized control arm (n = 41) received 7 days of IV amphotericin with flucytosine, then 7 days of fluconazole 1200 mg/d. The primary end point was cerebrospinal fluid (CSF) early fungicidal activity (EFA). RESULTS: We randomized 80 participants to oral LNC amphotericin + flucytosine with (n = 40) and without (n = 40) 2 IV loading doses and 41 control participants to IV amphotericin + flucytosine. Mean EFA was 0.40 log10 colony-forming units (CFU)/mL/d for all-oral LNC amphotericin, 0.42 log10  Cryptococcus CFU/mL/d for oral LNC amphotericin with IV loading doses, and 0.46 log10 CFU/mL/d for IV amphotericin controls. LNC amphotericin groups achieved 2-week CSF sterility in 63% (44 of 70) vs 68% (23 of 34) of controls. The 18-week survival was 85% (34 of 40) with all-oral LNC amphotericin, 90% (36 of 40) with oral LNC amphotericin given IV loading doses, and 85% (35 of 41) with IV amphotericin.Grade 3-4 laboratory adverse events occurred less frequently in LNC amphotericin groups (41%) than the IV amphotericin group (61%, P = .05), particularly for anemia (21% vs 44%; P = .01) and potassium (5% vs 17%; P = .04). CONCLUSIONS: This new oral amphotericin B LNC formulation appears promising for cryptococcal meningitis with antifungal activity, similar survival, and less toxicity than IV amphotericin. CLINICAL TRIALS REGISTRATION: NCT04031833.

Prevalence of Histoplasma Antigenuria among Outpatient Cohort with Advanced HIV in Kampala, Uganda.

In sub-Saharan Africa, an estimated 25% of people with HIV present with advanced HIV and are at high risk of opportunistic infections. Whereas histoplasmosis has occasionally been seen in Uganda, the understanding of the local risk of acute infection is limited. We sought to determine the prevalence of Histoplasma antigenuria using an enzyme immunoassay (EIA, clarus Histoplasma GM EIA, IMMY; Norman, OK, USA) in a cohort of outpatients with advanced HIV disease in Kampala, Uganda. Among the persons with positive urine Histoplasma antigen tests, we assessed their clinical presentation and outcomes. The EIA was run on stored urine samples as per the manufacturer's instructions. Specimens ≥1 EIA units were considered positive. Among the 388 tested urine samples, 4 (1.2%) were positive for Histoplasma antigen. The histoplasmosis prevalence among participants with a CD4 < 100 cells/mcL was 2.5% (4/158). Three of the four participants with a positive Histoplasma antigen test reported systemic symptoms consistent with histoplasmosis. All four participants had a positive urine lipoarabinomannan test and were treated for tuberculosis. By the four-week follow-up visit, all participants were clinically improved, alive, and in care without antifungal therapy. In advanced HIV, the clinical presentations of tuberculosis and histoplasmosis overlap. The value of histoplasmosis screening and pre-emptive treatment is an area of future research.

Two-Stage Adaptive Design for Prognostic Biomarker Signatures with a Survival Endpoint.

Cancer biomarker discoveries typically involve utilizing patient specimens. In practice, there is often strong desire to preserve high quality biospecimens for studies that are most likely to yield useful information. Previously, we proposed a two-stage adaptive design for binary endpoints which terminates the biomarker study in a futility interim if the model performance is unsatisfactory. In this work, we extend the two-stage design framework to accommodate time-to-event endpoints. The first stage of the procedure involves testing whether the measure of discrimination for survival models (C-index) exceeds a pre-specified threshold. We describe the computation of cross-validated C-index and evaluation of the statistical significance using re-sampling techniques. The second stage involves an independent model validation. Our simulation studies show that under the null hypothesis, the proposed design maintains type I error at the nominal level and has high probabilities of terminating the study early. Under the alternative hypothesis, power of the design is a function of the true event proportion, the sample size, and the targeted improvement in the discriminant measure. We apply the method to design of a prognostic biomarker study in patients with triple-negative breast cancer. Some practical aspects of the proposed method are discussed.

Return-to-baseline multiple imputation for missing values in clinical trials.

Return-to-baseline is an important method to impute missing values or unobserved potential outcomes when certain hypothetical strategies are used to handle intercurrent events in clinical trials. Current return-to-baseline approaches seen in literature and in practice inflate the variability of the "complete" dataset after imputation and lead to biased mean estimators when the probability of missingness depends on the observed baseline and/or postbaseline intermediate outcomes. In this article, we first provide a set of criteria a return-to-baseline imputation method should satisfy. Under this framework, we propose a novel return-to-baseline imputation method. Simulations show the completed data after the new imputation approach have the proper distribution, and the estimators based on the new imputation method outperform the traditional method in terms of both bias and variance, when missingness depends on the observed values. The new method can be implemented easily with the existing multiple imputation procedures in commonly used statistical packages.

Greater intraprocedural systolic blood pressure and blood pressure variability are associated with contrast-induced neurotoxicity after neurointerventional procedures.

BACKGROUND: Contrast-induced neurotoxicity (CIN) is a rare complication of neurointerventional procedures and its understanding remains limited. We evaluated the association of CIN with systemic hemodynamics in patients undergoing neuroendovascular interventions. METHODS: We conducted a 1:2 matched case-control study from a prospectively collected database of 2510 neurointerventional patients. We defined CIN as new neurological deficits presented ≤24 h post-operation after excluding other possible etiologies. We obtained demographic, clinical and imaging data, and baseline and intraprocedural blood pressures (BP) from medical records. The area between baseline and intraprocedural BP was used to measure sustained variability of BP over time. A generalized linear mixed model and generalized estimating equation were used to analyze the BP difference between groups over time. RESULTS: We evaluated 11 CIN cases and 22 controls. 2746 and 5837 min of continued BP data were analyzed for cases and controls, respectively. CIN cases had higher measurements and greater variability for: Systolic BP (SBP) [median 125 (IQR:121-147) vs. 114 (IQR:107-124) mmHg], median area above baseline [median 350 (IQR:25-1328) vs. 52 (IQR:0-293) mmHg*minutes] and mean arterial pressure (MAP) [median 85 (IQR:79-98) vs. 80 (IQR:74-89) mmHg]. CIN cases demonstrated a significant mean increase in SBP and MAP of 23.41 mmHg (p < 0.01) and 13.79 mmHg (p < 0.01) when compared to controls, respectively, over the perioperative time. CONCLUSION: Sustained hypertension and high BP variability may contribute to the pathophysiology of CIN. Acute hypertension can increase blood-brain barrier permeability and potentially allow contrast to leak into the brain parenchyma causing direct toxicity and CIN symptoms.

In-House Anesthesia and Interventional Radiology Technologist Support Optimize Mechanical Thrombectomy Workflow after Hours.

BACKGROUND AND PURPOSE: Prior literature suggests after-hours delay leads to poor functional outcomes in stroke patients undergoing thrombectomy. We aimed to evaluate the impact of time of presentation on mechanical thrombectomy (MT) metrics and its association with long-term functional outcome in an Interventional Radiology (IR) suite equipped operating room (OR) setting. METHODS: Retrospective review of prospectively maintained database on all stroke patients undergoing mechanical thrombectomy between January 2015 and December 2018 at our CSC. Work hours were defined by official OR work hours (Monday-Friday 7 AM and 5 PM) and after-hours as between 5 PM and 7 AM during weekdays and weekends as well as official hospital holidays. Primary outcome was 90-day modified Rankin Scale (mRS). Secondary outcomes included door to groin puncture time and procedural complications. RESULTS: A total of 315 patients were included in the analyses. 209 (66.4%) received mechanical thrombectomy after hours and 106 (33.6%) during work hours. There was no difference in the shift distribution of functional outcome on the mRS at 90 days (OR: 1.14, CI: 0.72-1.78, p=0.58) and the percentage of patients achieving functional independence (mRS 0-2) at 90 days (43.1% vs. 41.3%; p=0.83) between the after hour and work hour groups respectively. Similarly, there was no difference in median door to groin times and procedural complications among both groups, with significant year on year improvement in overall time metrics. CONCLUSIONS: Our study showed that undergoing MT during off-hours had similar functional outcomes when compared to MT during working hours in an OR setting. The after-hours deleterious effect might disappear when MT is performed in a system with 24-hours in-house Anesthesia and IR tech services.

Patterns of Care and Clinical Outcomes in Patients with Cerebral Sinus Venous Thrombosis.

OBJECTIVES: To explore the association between rurality, transfer patterns and level of care with clinical outcomes of CVST patients in a rural Midwestern state. MATERIALS AND METHODS: CVST patients admitted to the hospitals between 2005 and 2014 were identified by inpatient diagnosis codes from statewide administrative claims dataset. Records were linked across interhospital transfers using probabilistic linkage. Rurality was defined by Rural-Urban Commuting Areas using the 2-category approximation. Driving distances were estimated using GoogleMaps Application Programming Interface. Hospital stroke certification was defined by the Joint Commission. Severity of CVST was estimated by cost of care corrected for inflation and cost-to-charge ratios. Outcome was discharge disposition and total length of stay (LOS). Wilcoxon rank-sum, Chi-square, Fisher's exact tests and linear and logistic regressions were used. RESULTS: 168 CVST patients were identified (79.8% female; median age = 32, IQR = 24.0-45.5). Median LOS was four days (IQR = 2-7) and patients traveled a median of 8.1 miles (IQR = 2.5-28.5) to the first hospital; 42% of patients were transferred to a second hospital, 5% to a third. More than half (58.3%) bypassed the nearest hospital. 86% visit a primary or comprehensive stroke center (CSC) during their acute care. Rurality was not significantly associated with LOS or discharge disposition after adjusting for age, sex and cost of care. Patients in CSC demonstrated greater likelihood of being discharged home compared to at a primary stroke center after adjusting for age and disease severity (p = 0.008). CONCLUSIONS: While rurality was not significantly associated with LOS or disposition outcome, care at a CSC increases likelihood of being discharge home.

Association of admission clinical predictors and functional outcome in patients with Cerebral Venous and Dural Sinus Thrombosis.

OBJECTIVES: Cerebral venous sinus thrombosis (CVST) is a rare subtype of stroke that most commonly affects younger women. While most patients treated with anticoagulation therapy have good outcomes, a significant number go on to experience disability. The primary aim of this study was to identify objective, easily reproducible, clinical admission predictors of poor outcome at discharge in patients with CVST. PATIENTS AND METHODS: This was a retrospective cohort study of adult CVST patients admitted at our comprehensive stroke center between April 2004 and December 2017. The medical records of patients with a CVST discharge diagnosis code were reviewed for diagnosis confirmation and extraction of clinical and demographic admission data. Multivariable logistic regression was used to build predictive models of objective, standardized examination signs and adjusted for confounders. The primary endpoint was modified Rankin Scale score at discharge defined as good outcome (0-2) and poor outcome (3-6). Mortality was the secondary endpoint. RESULTS: A total of 176 CVST patients were identified. Most patients were white (91 %) and female (65 %). The median age was 40 years old. Headache was the most commonly reported symptom (74 %). Intracranial hemorrhage (ICH) was present in 27 % of patients, venous infarct occurred in 22 % of the patients, and 12 % had both. Age (OR = 1.03, 95 % CI 1.01-1.05), abnormal level of consciousness (OR = 4.38, 95 % CI 1.86-8.88), and focal motor deficits (OR = 3.49, 95 % CI 1.49-8.15) were found to be predictive of poor functional outcome. Pre-hospitalization infections (OR = 5.22, 95 % CI 1.51-18.07) and abnormal level of consciousness (OR = 9.22, 95 % CI 2.34-36.40) were significant predictors of mortality. The predictive effect remained significant after adjusting by median PTT level, presence of intracranial hemorrhage, and venous infarct. CONCLUSIONS: Age, abnormal level of consciousness, and focal motor deficits identified at admission are independently associated with poor outcome in CVST patients. These frequently prevalent, easily reproducible examination signs represent the first step to develop a clinical prediction tool toward stratifying CVST patients with poor prognosis at admission.

Computed Tomography Angiogram Derived From Computed Tomography Perfusion Done with Low Iodine Volume Protocol Preserves Diagnostic Yield for Middle Cerebral Artery-M2 Occlusions.

BACKGROUND: Computed tomography angiogram (CTA) derived from computed tomography perfusion (CTP) has been proposed to avoid addition of separate CT perfusion protocol for selection of large vessel occlusion in acute stroke patients. Previous studies have validated this technique for proximal large vessel occlusions. In this study, we test reliability for identifying M2 occlusions on CTA derived from CTP. METHODS: Through a retrospective search of the institutional thrombectomy database, we identified 28 cases with M2-MCA occlusion, of which 24 met the inclusion criteria for analysis. An additional 20 cases without M2-MCA occlusion (either normal or M1-MCA occlusion) were randomly mixed in the database to reduce observer bias. The baseline images of the CTP study in these 48 cases were then independently analyzed by 3 readers with varying level of expertise. The digital subtraction angiography (DSA) images were also independently reviewed where available. The percentage of agreement among reviewers as well as the probability of agreement of the reviewers, when compared to the DSA findings was also calculated. RESULTS: The observed agreement for the image quality amongst the 3 readers (n = 48) varied between 0.78 and 0.95 and tended to be higher for the M1 segment MCA and lower for distal M2-MCA. The observed agreements comparing 3 image reviewers versus DSA in M2 patients (n = 24) was 98% for identifying occlusion (95% CI 95%-100%), 94% for identifying proximal M2 occlusion (95% CI 88%-98%), and 91% (95% CI 84%-97%) and 90% (95% CI 83%-95%), respectively for correctly identifying inferior and superior branch of M2 occlusion. CONCLUSION: CTA data derived from CT Perfusion study preserves diagnostic yield for correctly identifying M2 occlusion.

Stroke mechanisms and outcomes of isolated symptomatic basilar artery stenosis.

BACKGROUND: While diffuse atherosclerotic disease affecting the posterior circulation has been described extensively, the prevalence, natural history and angiographic characteristics of isolated symptomatic basilar artery stenosis (ISBAS) remain unknown. METHODS: We reviewed our prospective institutional database to identify patients with ≥50% symptomatic basilar artery (BA) stenosis without significant atherosclerotic burden in the vertebral or posterior cerebral arteries. Stroke mechanism, collateral circulation, and degree and length of stenosis were analysed. The primary outcome was time from index event to new transient ischaemic attack (TIA), acute ischaemic stroke (AIS) or death. Other outcome variables included modified Rankin Scale (mRS) score on discharge and last follow-up. RESULTS: Of 6369 patients with AIS/TIA, 91 (1.43%) had ISBAS. Seventy-three (80.2%) patients presented with AIS and 18 (19.8%) with TIA. Twenty-nine (31.9%) were women and the median age was 66.8±13.6 years. The mean follow-up time was 2.7 years. The most common stroke mechanism was artery-to-artery thromboembolism (45.2%), followed by perforator occlusion (28.7%) and flow-dependent/hypoperfusion (15.1%). The percentage of stenosis was lower in patients who had favourable outcome compared with those with mRS 3-6 on discharge (78.3±14.3 vs 86.9±14.5, p=0.007). Kaplan-Meier curves showed higher recurrence/death rates in patients with ≥80% stenosis, mid-basilar location and poor collateral circulation. Approximately 13% of patients with ISBAS presented with complete BA occlusion. CONCLUSION: ISBAS is an uncommon (1.43%) cause of TIA and AIS. Men in their 60s are mostly affected, and artery-to-artery embolism is the most common stroke mechanism. Mid-basilar location, ≥80% stenosis and poor collateral circulation are important factors associated with worse prognosis.