The role of inflammation and pain in osteoarthritis (OA) is not fully understood. We evaluated the association between pro-inflammatory biomarkers and pain. We used baseline data and samples from a randomized controlled trial of colchicine for symptomatic knee OA. Severity of pain of the more symptomatic knee was assessed by National Health and Nutrition Examination Survey-I (NHANES-I) criterion and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain index. Pains on movement and at rest were self-reported on an 11-point Likert scale. Severity of radiographic tibiofemoral OA was assessed by Kellgren and Lawrence (KL) grade. Concentrations of synovial fluid (sf) IL-1β, IL-6, IL-8, TNFα, C-terminal telopeptides of Type I collagen (CTXI) and C-telopeptide of Type II collagen (CTXII), as well as urinary (u) CTXII were measured. Of the 109 patients enrolled in the study, 70 patients (70% women) with synovial fluid obtained by direct aspiration were included for analysis. The mean ± SD age and body mass index (BMI) of the patients were 57.6 ± 8.3 years and 28.8 ± 5.2 kg/m2. After adjustment for age, sex, and BMI, sf IL-6 and IL-8 were statistically significantly associated with 11-point pain on movement, but not with pain at rest. No significant associations were observed with WOMAC pain scores. sf IL-1β (analyzed as detectable/non-detectable) was inversely associated with pain. In contrast, after adjustment, Sf TNFα was associated with WOMAC total pain and both pain on movement and at rest. sf/u CTXII was associated with radiographic severity, but not with knee pain. This study provides indication that OA pain mechanisms may differ according to the characteristics of the pain.

Uric acid may activate an innate immune response in osteoarthritis (OA), contributing to disease pathology and progression. We evaluated the effectiveness of colchicine on pain and function in symptomatic knee OA (KOA) and the underlying mechanism of action. Colchicine effectiveness in symptoms and inflammation modification in knee osteoarthritis (COLKOA) was a double-blind, placebo-controlled, randomized trial comparing 16 weeks of treatment with 0.5 mg twice-daily oral colchicine to placebo for knee osteoarthritis (KOA). The primary endpoint was ≥30% improvement in total Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) score at week 16. Secondary endpoints included improvement in pain (0-10 Likert scales); WOMAC pain; patient global assessment (0-100); physical function; the OARSI-OMERACT response; quality of life; and change in serum, urine, synovial fluid (SF) biomarkers of cartilage metabolism and inflammation, and plasma/SF colchicine concentrations. Of 109 randomly assigned participants, 39% (95% confidence interval (CI) 27-52%) and 49% (95% CI 36-62%) in the colchicine and placebo arms respectively met the primary endpoint at study end (P = 0.284, odds ratio 0.66, 95% CI 0.31-1.41). No strong evidence of treatment differences was identified on clinical secondary endpoints. Treatment significantly reduced mean serum hs-CRP (P = 0.008) and SF CTXI (P = 0.002); treatment tended to reduce inflammatory markers (SF IL-6, IL8, TNFα, CD14 and IL-18), but these differences were not statistically significant. Colchicine (0.5 mg twice-daily orally) reduced inflammation and high bone turnover biomarkers known to be associated with OA severity and progression risk, but did not reduce KOA symptoms over a 16-week study period. A longer-term study to evaluate for slow-acting disease modifying effects is warranted. The trial has been registered at clinicaltrials.gov as NCT02176460. Date of registration: June 26, 2014.

Scientists and engineers commonly use simulation models to study real systems for which actual experimentation is costly, difficult, or impossible. Many simulations are stochastic in the sense that repeated runs with the same input configuration will result in different outputs. For expensive or time-consuming simulations, stochastic kriging is commonly used to generate predictions for simulation model outputs subject to uncertainty due to both function approximation and stochastic variation. Here, we develop and justify a few guidelines for experimental design, which ensure accuracy of stochastic kriging emulators. We decompose error in stochastic kriging predictions into nominal, numeric, parameter estimation, and parameter estimation numeric components and provide means to control each in terms of properties of the underlying experimental design. The design properties implied for each source of error are weakly conflicting and broad principles are proposed. In brief, space-filling properties, “small fill distance” and “large separation distance,” should be balanced with replication at distinct input configurations, with number of replications depending on the relative magnitudes of stochastic and process variability. Nonstationarity implies higher input density in more active regions, while regression functions imply a balance with traditional design properties. A few examples are presented to illustrate the results. Supplementary materials providing proofs of the theoretical results and code for comparisons are available online.

Evidence has suggested survival benefits of maintenance for advanced NSCLC patients not progressing after first‐line chemotherapy. Additionally, particular first‐line targeted therapies have shown survival improvements in selected populations. Optimal first‐line and maintenance therapies remain unclear. Here, currently available evidence was synthesized to elucidate optimal first‐line and maintenance therapy within patient groups. Literature was searched for randomized trials evaluating first‐line and maintenance regimens in advanced NSCLC patients. Bayesian network meta‐analysis was performed within molecularly and clinically selected groups. The primary outcome was combined clinically meaningful OS and PFS benefits. A total of 87 records on 56 trials evaluating first‐line treatments with maintenance were included. Results showed combined clinically meaningful OS and PFS benefits with particular first‐line with maintenance treatments, (1) first‐line intercalated chemotherapy+erlotinib, maintenance erlotinib in patients with EGFR mutations, (2) first‐line afatinib, maintenance afatinib in patients with EGFR deletion 19, (3) first‐line chemotherapy + bevacizumab, maintenance bevacizumab in EGFR wild‐type patients, (4) chemotherapy+conatumumab, maintenance conatumumab in patients with squamous histology, (5) chemotherapy+cetuximab, maintenance cetuximab or chemotherapy + necitumumab, maintenance necitumumab in EGFR FISH‐positive patients with squamous histology, and (6) first‐line chemotherapy+bevacizumab, maintenance bevacizumab or first‐line sequential chemotherapy+gefitinib, maintenance gefitinib in patients clinically enriched for EGFR mutations with nonsquamous histology. No treatment showed combined clinically meaningful OS and PFS benefits in patients with EGFR L858R or nonsquamous histology. Particular first‐line with maintenance treatments show meaningful OS and PFS benefits in patients selected by EGFR mutation or histology. Further research is needed to achieve effective therapy for patients with EGFR mutation L858R or nonsquamous histology.

The Gaussian process is a standard tool for building emulators for both deterministic and stochastic computer experiments. However, application of Gaussian process models is greatly limited in practice, particularly for large-scale and many-input computer experiments that have become typical. We propose a multi-resolution functional ANOVA model as a computationally feasible emulation alternative. More generally, this model can be used for large-scale and many-input non-linear regression problems. An overlapping group lasso approach is used for estimation, ensuring computational feasibility in a large-scale and many-input setting. New results on consistency and inference for the (potentially overlapping) group lasso in a high-dimensional setting are developed and applied to the proposed multi-resolution functional ANOVA model. Importantly, these results allow us to quantify the uncertainty in our predictions. Numerical examples demonstrate that the proposed model enjoys marked computational advantages. Data capabilities, both in terms of sample size and dimension, meet or exceed best available emulation tools while meeting or exceeding emulation accuracy.