Search

Background and Aim: Patients with liver cirrhosis often face a grave threat from infected ascites (IA). However, a well‐established prognostic model for this complication has not been established in routine clinical practice. Therefore, we aimed to assess mortality risk in patients with liver cirrhosis and IA. Methods: We conducted a retrospective study across three tertiary hospitals, enrolling 534 adult patients with cirrhotic liver and IA, comprising 465 with spontaneous bacterial peritonitis (SBP), 34 with bacterascites (BA), and 35 with secondary peritonitis (SP). To determine the attributable mortality risk linked to IA, these patients were matched with 122 patients with hydropic decompensated liver cirrhosis but without IA. Clinical, laboratory, and microbiological parameters were assessed for their relation to mortality using univariable analyses and a multivariable random forest model (RFM). Least absolute shrinkage and selection operator (Lasso) regression model was used to establish an easy‐to‐use mortality prediction score. Results: The in‐hospital mortality risk was highest for SP (39.0%), followed by SBP (26.0%) and BA (25.0%). Besides illness severity markers, microbiological parameters, such as Candida spp., were identified as the most significant indicators for mortality. The Lasso model determined 15 parameters with corresponding scores, yielding good discriminatory power (area under the receiver operating characteristics curve = 0.89). Counting from 0 to 83, scores of 20, 40, 60, and 80 corresponded to in‐hospital mortalities of 3.3%, 30.8%, 85.2%, and 98.7%, respectively. Conclusion: We developed a promising mortality prediction score for IA, highlighting the importance of microbiological parameters in conjunction with illness severity for assessing patient outcomes. [ABSTRACT FROM AUTHOR]

Background: Despite advances in coronary revascularization and in heart failure management, myocardial infarction survivors remain at substantially increased mortality risk. Precise risk assessment and risk‐adapted follow‐up care are crucial to improve their outcomes. Recently, the fragmented QRS complex, i.e. the presence of additional spikes within the QRS complexes on a 12‐lead electrocardiogram, has been discussed as a potential non‐invasive risk predictor in cardiac patients. Hypothesis: The aim of this study was to evaluate the prognostic meaning of the fragmented QRS complex in myocardial infarction survivors. Methods: 609 patients with narrow QRS complexes <120 ms were included in a prospective cohort study while hospitalized for myocardial infarction and followed for 5 years. Results: The prevalence of the fragmented QRS complex in these patients amounted to 46.8% (285 patients). These patients had no increased hazard of all‐cause death (HR 0.84, 95%–CI 0.45–1.57, p = 0.582) with a mortality rate of 6.0% compared to 7.1% in patients without QRS fragmentations. Furthermore, the risks of cardiac death (HR 1.28, 95%–CI 0.49–3.31, p = 0.613) and of non‐cardiac death (HR 0.6, 95%–CI 0.26–1.43, p = 0.25) were not significantly different in patients with QRS fragmentations. However, patients with QRS fragmentations had increased serum creatine kinase concentrations (1438U/l vs. 1160U/l, p = 0.039) and reduced left ventricular ejection fractions (52% vs. 54%, p = 0.011). Conclusions: The hypothesis that QRS fragmentation might be a prognostic parameter in survivors of myocardial infarction was not confirmed. But those with QRS fragmentation had larger myocardial infarctions, as measured by creatine kinase and left ventricular ejection fraction. [ABSTRACT FROM AUTHOR]

Tumour budding (TB) has been associated with adverse clinicopathological factors and poor survival in a plethora of therapy‐naïve carcinoma entities including gastric adenocarcinoma (GC). As conventional histopathological grading is usually omitted in the post‐neoadjuvant setting of GC, our study aimed to investigate the prognostic impact of TB in GCs resected after neoadjuvant therapy. We evaluated TB according to the criteria from the International Tumour Budding Consensus Conference (ITBCC) in 167 post‐neoadjuvant resections of intestinal‐type GC and correlated the results with overall survival (OS) and clinicopathological parameters. GCs were categorised into Bd1 (0–4 buds, low TB), Bd2 (5–9 buds, intermediate TB), and Bd3 (≥10 buds, high TB). Carcinomas with intermediate and high TB were significantly enriched in higher ypTNM stages and strongly associated with reduced 5‐year OS in univariable analyses (p < 0.001). In multivariable analyses including sex, age, resection status, UICC stage, and tumour regression grading, TB remained a stage‐independent predictor of survival (p < 0.001, hazard ratio Bd2: 2.60, Bd3: 4.74). The assessment of TB according to the ITBCC criteria provides valuable prognostic information in the post‐neoadjuvant setting of intestinal‐type GC and may be a considerable substitute for the conventional grading system in GCs after neoadjuvant therapy. [ABSTRACT FROM AUTHOR]

We investigated the clinical impact of elevated microsatellite instability at selected tetranucleotide (EMAST) repeats in the context of neoadjuvant chemotherapy (CTx) in gastric/gastro‐oesophageal adenocarcinomas. We analysed 583 resected tumours (272 without and 311 after CTx) and 142 tumour biopsies before CTx. If at least two or three of the five tetranucleotide repeat markers tested showed instability, the tumours were defined as EMAST (2+) or EMAST (3+), respectively. Expression of mismatch repair proteins including MSH3 was analysed using immunohistochemistry. Microsatellite instability (MSI) and Epstein–Barr virus (EBV) positivity were determined using standard assays. EMAST (2+) and (3+) were detected in 17.8 and 11.5% of the tumours, respectively. The frequency of EMAST (2+) or (3+) in MSI‐high (MSI‐H) tumours was 96.2 or 92.5%, respectively, demonstrating a high overlap with this molecular subtype, and the association of EMAST and MSI status was significant (each overall p < 0.001). EMAST (2+ or 3+) alone in MSI‐H and EBV‐negative tumours demonstrated only a statistically significant association of EMAST (2+) positivity and negative lymph node status (42.3% in EMAST (2+) and 28.8% in EMAST negative, p = 0.045). EMAST alone by neither definition was significantly associated with overall survival (OS) of the patients. The median OS for EMAST (2+) patients was 40.0 months (95% confidence interval [CI] 16.4–63.6) compared with 38.7 months (95% CI 26.3–51.1) for the EMAST‐negative group (p = 0.880). The median OS for EMAST (3+) patients was 46.7 months (95% CI 18.2–75.2) and 38.7 months (95% CI 26.2–51.2) for the negative group (p = 0.879). No statistically significant association with response to neoadjuvant CTx was observed (p = 0.992 and p = 0.433 for EMAST (2+) and (3+), respectively). In conclusion, our results demonstrate a nearly complete intersection between MSI‐H and EMAST and they indicate that EMAST alone is not a distinct instability type associated with noticeable clinico‐pathological characteristics of gastric carcinoma patients. [ABSTRACT FROM AUTHOR]

Introduction: In mice there might be an association between sleep deprivation and amyloid β plasma levels. Hence, we examined whether amyloid plasma levels are associated with sleep duration or fragmentation in 17 psychiatrists on-call. Methods: Amyloid β (Aβ)42, Aβ40, and soluble amyloid precursor protein β (sAPP-β) plasma concentrations were measured at the beginning and end of 90 on-call nights, and analyzed using generalized linear models. Results: In on-call nights, a 10.7% reduction of Aβ42 was revealed overnight. Every single short sleep interruption diminished this reduction by 5.4%, as well as every pg/mL of sAPP-β by 1.2%, each copy of APOE ε4 by 10.6%, and each year of professional experience by 3.0%. Discussion: The extent of sleep fragmentation diminishes the physiological overnight reduction of Aβ42 but not Aβ40 plasma levels in the same direction as the enzyme for Aβ42 production, the genetic risk factor for Alzheimer's disease (AD), and on-call experience. Might on-call duty and sleep fragmentation in general alter the risk for AD? [ABSTRACT FROM AUTHOR]

The goal in stratified medicine is to administer the "best" treatment to a patient. Not all patients might benefit from the same treatment; the choice of best treatment can depend on certain patient characteristics. In this article, it is assumed that a time-to-event outcome is considered as a patient-relevant outcome and a qualitative interaction between a continuous covariate and treatment exists, ie, that patients with different values of one specific covariate should be treated differently. We suggest and investigate different methods for confidence interval estimation for the covariate value, where the treatment recommendation should be changed based on data collected in a randomized clinical trial. An adaptation of Fieller's theorem, the delta method, and different bootstrap approaches (normal, percentile-based, wild bootstrap) are investigated and compared in a simulation study. Extensions to multivariable problems are presented and evaluated. We observed appropriate confidence interval coverage following Fieller's theorem irrespective of sample size but at the cost of very wide or even infinite confidence intervals. The delta method and the wild bootstrap approach provided the smallest intervals but inadequate coverage for small to moderate event numbers, also depending on the location of the true changepoint. For the percentile-based bootstrap, wide intervals were observed, and it was slightly conservative regarding coverage, whereas the normal bootstrap did not provide acceptable results for many scenarios. The described methods were also applied to data from a randomized clinical trial comparing two treatments for patients with symptomatic, severe carotid artery stenosis, considering patient's age as predictive marker. [ABSTRACT FROM AUTHOR]

Background and Aim: In contrast to the first peak of multi‐organ failure in acute pancreatitis, the second peak is mostly triggered by septic complications. Our aim was to analyze the spectrum of pathogens and antimicrobial resistance development in relation to the time‐course of the disease and its clinical outcome. Methods: One hundred twenty‐two patients with acute necrotizing pancreatitis undergoing pancreas puncture at two tertiary academic medical centers in Germany were retrospectively analyzed. Results: At species level, there was a change in spectrum from Enterococcus faecalis (∆d150 − d1 = 14.6% − 16.7% = −2.1%) to Enterococcus faecium (∆d150 − d1 = 93.1% − 16.3% = 76.8%) (P < 0.001) and from Candida albicans (∆d150 − d1 = 39.7% − 23.6% = 16.1%) to non‐albicans Candida spp. (∆d150 − d1 = 43.5% − 6.4% = 37.1%) (P = 0.005). Time‐to‐event analysis of acquired antimicrobial resistance showed that the overall number of patients with Enterobacteriaceae presented an antimicrobial susceptibility decrease by 59.7% (∆d1 − d100 = 87.0% − 27.3% = 59.7%). The cumulative incidence of multi‐resistant bacteria increased with length of hospital stay (∆d150 − d1 = 49.1% − 3.1% = 46.0%) (P = 0.004). Multivariable logistic regression analysis in relation to the pathogen spectrum and antimicrobial resistance development showed a significantly higher mortality for non‐albicans Candida spp. (P = 0.039, odds ratio [OR] = 3.32 [95% confidence interval [CI]: 1.07–10.35]), E. faecium (P = 0.009, OR = 3.73 [95% CI: 1.38–10.05]), and multi‐resistant bacteria (P = 0.007, OR = 5.08 [95% CI: 1.55–16.66]). Conclusions: Antimicrobial treatment of infected pancreatic necrosis becomes more challenging over time, owing to a change in spectrum favoring difficult‐to‐treat pathogens and an increase in multi‐resistant bacteria associated with worse clinical outcomes (World Health Organization trial registration number: DRKS00014785). [ABSTRACT FROM AUTHOR]

Epstein–Barr virus positivity (EBV(+)) and high‐microsatellite instability (MSI‐H) have been identified as molecular subgroups in gastric carcinoma. The aim of our study was to determine the prognostic and predictive relevance of these subgroups in the context of platinum/5‐fluorouracil (5‐FU) based preoperative chemotherapy (CTx). Additionally, we investigated the clinical relevance of the low‐MSI (MSI‐L) phenotype. We analysed 760 adenocarcinomas of the stomach or the gastro‐oesophageal junction encompassing 143 biopsies before CTx and 617 resected tumours (291 without and 326 after CTx). EBV was determined by PCR and in situ hybridisation for selected cases. MSI was analysed by PCR using five microsatellite markers and classified as MSI‐H and MSI‐L. Frequencies of EBV(+), MSI‐H and MSI‐L in the biopsies before CTx were 4.2, 10.5 and 4.9% respectively. EBV(+) or MSI‐H did not correlate with response, but MSI‐L was associated with better response (p = 0.011). In the resected tumours, frequencies of EBV(+), MSI‐H and MSI‐L were 3.9, 9.6 and 4.5% respectively. Overall survival (OS) was significantly different in the non‐CTx group (p = 0.014). Patients with EBV(+) tumours showed the best OS, followed by MSI‐H. MSI‐L was significantly associated with worse OS (hazard ratio [HR], 2.21; 95% confidence interval [CI], 1.21–4.04, p = 0.01). In the resected tumours after CTx, MSI‐H was also associated with increased OS (HR, 0.54; 95% CI, 0.26–1.09, p = 0.085). In multivariable analysis, molecular classification was an independent prognostic factor in the completely resected (R0) non‐CTx group (p = 0.035). In conclusion, MSI‐H and EBV(+) are not predictive of response to neoadjuvant platinum/5‐FU based CTx, but they are indicative of a good prognosis. In particular, MSI‐H indicates a favourable prognosis irrespective of treatment with CTx. MSI‐L predicts good response to CTx and its negative prognostic effect for patients treated with surgery alone suggests that MSI‐L might help to identify patients with potentially high‐benefit from preoperative CTx. [ABSTRACT FROM AUTHOR]

Ideally, prediction rules should be published in such a way that readers may apply them, for example, to make predictions for their own data. While this is straightforward for simple prediction rules, such as those based on the logistic regression model, this is much more difficult for complex prediction rules derived by machine learning tools. We conducted a survey of articles reporting prediction rules that were constructed using the random forest algorithm and published in PLOS ONE in 2014–2015 in the field "medical and health sciences", with the aim of identifying issues related to their applicability. Making a prediction rule reproducible is a possible way to ensure that it is applicable; thus reproducibility is also examined in our survey. The presented prediction rules were applicable in only 2 of 30 identified papers, while for further eight prediction rules it was possible to obtain the necessary information by contacting the authors. Various problems, such as nonresponse of the authors, hampered the applicability of prediction rules in the other cases. Based on our experiences from this illustrative survey, we formulate a set of recommendations for authors who aim to make complex prediction rules applicable for readers. All data including the description of the considered studies and analysis codes are available as supplementary materials. [ABSTRACT FROM AUTHOR]

Abstract: Background and Aim: Mortality of cirrhotic patients after emergency care admission is high, and prognostic factors can help in prioritizing patients. The aim of our study was to assess the association between levels of cardiac troponin T (cTnT) and 1‐year mortality in patients with liver cirrhosis without known cardiac disease, who were admitted to the emergency department (ED). Methods: All patients with cirrhosis presented to the ED from October 2009 until August 2015 who had an initial cTnT value measured with the first lab panel were retrospectively analyzed with a follow‐up of 365 days. Results: Of a total of 237 cirrhotic ED patients, cTnT measurements were available for 87 (63% men, mean age 58.9 ± 11.0 years, and median Model for End‐stage Liver Disease score was 15 [25th–75th percentile: 10–19]). Chronic Liver Failure Consortium acute‐on‐chronic liver failure (CLIF‐C‐ACLF) score was 33. Forty‐three patients (49%) had cTnT values above the normal range (14 ng/L), of which 19 (22%) had values over 30 ng/L. Two patients were lost to follow‐up. In multivariable analysis, both CLIF‐C‐ACLF (hazard ratio 1.072 per point increase; 95% confidence interval 1.029–1.117; P < 0.001) and cTnT (hazard ratio 1.014 per ng/L increase; 95% confidence interval 1.004–1.024; P = 0.008) emerged as independently associated with mortality. Conclusions: A large proportion of cirrhotic patients in the ED have elevated levels of cTnT even if there is no evidence of cardiac disease. Elevated cTnT is associated with increased mortality during 1 year after correcting for Model for End‐stage Liver Disease and CLIF‐C‐ACLF scores. [ABSTRACT FROM AUTHOR]

Cefuroxime is widely used for antibiotic prophylaxis in orthopedic surgery. However, a recent study indicated a dose-dependent reduction in osteoblast function in vitro. Nevertheless, cell culture might not sufficiently imitate the complex process of bone remodeling. As data concerning possible in vivo interactions of cefuroxime on fracture healing are completely missing, we investigated the following hypothesis: Does Cefuroxime impair bone healing in vivo? Therefore, 34 male Wistar rats were randomised to cefuroxime-treated or control groups, a Kirschner wire was inserted into right femora and closed transverse fractures were produced. Twenty-one days later, the structural properties of the fracture callus in the early fracture healing phase were evaluated via a combination of micro-CT (μCT), biomechanics and histology. µCT demonstrated similar values in the cefuroxime and control group (e.g., callus volume: 67.19 ± 14.90 mm3 vs. 55.35 ± 6.74 mm3, p = 0.12; density: 635.48 ± 14.81 mg HA/cm3 vs. 647.87 ± 13.01 mg HA/cm3, p = 0.16). Biomechanically, similar values were again determined between the groups, in terms of both maximum load (77.65 ± 41.82 vs. 78.54 ± 20.52, p = 0.95) and stiffness (122.44 ± 81.16 vs. 123.74 ± 60.08, p = 0.97). Histological findings were consistent with the radiographic results. Thus, no relevant differences between the cefuroxime and control groups could be found and the reported negative effects on osteoblasts in vitro were not confirmed in vivo by using standard concentrations of cefuroxime. In conclusion, cefuroxime can reasonably be recommended in a clinical setting as an antibiotic therapy when fracture healing is involved. However, supraphysiological doses were not evaluated, which may be present when cefuroxime is used as an additive to bone cement and released over time. Therefore, future studies should evaluate the in vivo effects of prolonged high cefuroxime doses on implant incorporation. © 2016 Wiley Periodicals, Inc. J Biomed Mater Res Part B: Appl Biomater, 105B: 2282-2291, 2017. [ABSTRACT FROM AUTHOR]

Two nonparametric methods for the identification of subgroups with outstanding outcome values are described and compared to each other in a simulation study and an application to clinical data. The Patient Rule Induction Method (PRIM) searches for box-shaped areas in the given data which exceed a minimal size and average outcome. This is achieved via a combination of iterative peeling and pasting steps, where small fractions of the data are removed or added to the current box. As an alternative, Classification and Regression Trees (CART) prediction models perform sequential binary splits of the data to produce subsets which can be interpreted as subgroups of heterogeneous outcome. PRIM and CART were compared in a simulation study to investigate their strengths and weaknesses under various data settings, taking different performance measures into account. PRIM was shown to be superior in rather complex settings such as those with few observations, a smaller signal-to-noise ratio, and more than one subgroup. CART showed the best performance in simpler situations. A practical application of the two methods was illustrated using a clinical data set. For this application, both methods produced similar results but the higher amount of user involvement of PRIM became apparent. PRIM can be flexibly tuned by the user, whereas CART, although simpler to implement, is rather static. [ABSTRACT FROM AUTHOR]

Background Perioperative acquired factor XIII deficiency has been looked upon as a potential cause of postoperative bleeding in adult cardiac surgery. Methods Forty-four infants were prospectively studied for the time course of factor XIII in plasma and the effect on chest tube drainage ( CTD) and transfusion requirements in the first 24 h after surgery. A reconstituted blood prime ( RBP) with fresh-frozen plasma ( FFP) and packed red blood cells ( PRBC) was used. Samples were taken at baseline, after cardiopulmonary bypass and upon arrival in the ICU. Differences in blood loss and transfusion requirements based on a cutoff value of 70% factor XIII activity at the time of ICU admission were also calculated. Results Baseline factor XIII activity was 79%, decreased to 71% after CPB ( P = 0.102) and increased back up to 77% at ICU arrival ( P = 0.708). There was no significant correlation between factor XIII, CTD, age, cyanosis, platelet count, and transfusion requirements at any time point. Only preoperative fibrinogen levels correlated significantly with factor XIII activity. Perioperative blood transfusions ( PRBC P = 0.712, FFP P = 0.909, platelets P = 0.807) and chest tube losses ( P = 0.424 at 6 h and P = 0.215 at 24 h) were not significantly different above or below a 70% factor XIII activity at ICU arrival. Conclusion Factor XIII activity in infants with congenital heart defects is within the lower range of normal adults, independent of patient's age and the presence of cyanosis. Reconstituted blood prime maintains factor XIII activity at sufficient levels during pediatric cardiac surgery. We could not detect a correlation between FXIII and CTD. [ABSTRACT FROM AUTHOR]

Background: The aim of this study was to compare single-port access cholecystectomy (SPA) with the standard laparoscopic technique (LC) regarding the duration of the operation, complications, learning curve, late postoperative quality of life (QoL) and the incidence of incisional hernias. Methods: Between June 2009 and December 2011, a total of 122 SPA cholecystectomies were performed in our hospital. Simultaneously, 310 patients were operated on with the LC technique. In the LC group, 100 patients met the same criteria defined for SPA surgery. The two groups (SPA and LC) were compared by multivariable regression analysis. Endpoints of this study were quality of life (QoL) after 6 months by the EQ-5D questionnaire 5L and the incidence of incisional hernia 1 year after surgery. Operating time, hospital stay, and perioperative complications were also measured and compared. The median follow-up was 9.2 months (3-25 months). Results: The patients in the SPA group were younger and more often female. The mean operating time for group SPA was 73 min (35-136 min)-significantly longer than that for group LC with 60 min (33-190 min) ( p < 0.001). Additional trocars were used in 8 of 122 (6.5 %) SPA patients. A conversion to open cholecystectomy was not necessary in SPA patients. The conversion rate in the LC group to open cholecystectomy was 2 % (2/100). The perioperative and postoperative complications and incisional hernia (5.5 %) were the same in both groups. QoL was significantly better in the SPA group in terms of mobility ( p = 0,002), usual activity ( p = 0.036), and overall anxiety ( p = 0.026). Conclusions: SPA cholecystectomy is safe, although the operation is significantly longer. No differences in terms of major complications or the incidence of incisional hernia were seen after 1 year. QoL was significantly better in patients operated on with the SPA technique. [ABSTRACT FROM AUTHOR]

Objective: Bacterial translocation, causing intestinal inflammation, is one of the key mechanisms in the pathogenesis of hepatic encephalopathy (HE) and spontaneous bacterial peritonitis (SBP) The presence of fecal calprotectin quantitatively relates to intestinal neutrophil migration and is therefore considered as a marker of intestinal inflammation. We aimed to assess the role of fecal calprotectin concentrations (FCCs) in diagnosing the onset and severity of HE and SBP. Methods: Sixty-one cirrhotics were prospectively included. Forty-two subjects served as controls. Several complications of cirrhosis were diagnosed by reference methods. Stool samples were collected for measuring FCCs. Patients revealing other causes of abnormal calprotectin results, e.g. gastrointestinal bleeding or inflammatory bowel disease were excluded. Multivariate analysis of cirrhosis-associated complications and their relation to FCCs was performed. Results: Fecal calprotectin concentrations were higher in cirrhotics compared with controls ( P<0.001). Among cirrhotics, FCCs were elevated dependent on the severity of liver disease as assessed by Child- and model for end-stage liver disease-scores. The corresponding correlation co-efficients by Spearman's were 0.577 ( P<0.001) and 0.303 ( P=0.018) respectively. A correlation emerged between elevated FCCs and HE grading as measured by West-Haven criteria and critical flicker frequency (both P<0.001; sensitivity=0.94 and 0.93, specificity=0.95 and 0.89 respectively) and SBP ( P<0.02; sensitivity=0.71, specificity=0.79). FCCs were higher in cirrhotic subjects with additional extra-intestinal inflammation ( P<0.01; sensitivity=0.65, specificity=0.8). The Pearsons correlation coefficients were 0.190 and 0.164 revealing no influence ( P=0.142 and P=0.207) of laboratory parameters of systemic inflammation on FCCs in cirrhotic subgroup. Conclusions: Fecal calprotectin concentrations serve as a screening tool for HE and SBP. Assessment of FCCs may faciliate grading of HE-severity. [ABSTRACT FROM AUTHOR]

Identification of relevant biomarkers that are associated with a treatment effect is one requirement for adequate treatment stratification and consequently to improve health care by administering the best available treatment to an individual patient. Various statistical approaches were proposed that allow assessing the interaction between a continuous covariate and treatment. Nevertheless, categorization of a continuous covariate, e.g., by splitting the data at the observed median value, appears to be very prevalent in practice. In this article, we present a simulation study considering data as observed in a randomized clinical trial with a time-to-event outcome performed to compare properties of such approaches, namely, Cox regression with linear interaction, Multivariable Fractional Polynomials for Interaction (MFPI), Local Partial-Likelihood Bootstrap (LPLB), and the Subpopulation Treatment Effect Pattern Plot (STEPP) method, and of strategies based on categorization of continuous covariates (splitting the covariate at the median, splitting at quartiles, and using an "optimal" split by maximizing a corresponding test statistic). In different scenarios with no interactions, linear interactions or nonlinear interactions, type I error probability and the power for detection of a true covariate-treatment interaction were estimated. The Cox regression approach was more efficient than the other methods for scenarios with monotonous interactions, especially when the number of observed events was small to moderate. When patterns of the biomarker-treatment interaction effect were more complex, MFPI and LPLB performed well compared to the other approaches. Categorization of data generally led to a loss of power, but for very complex patterns, splitting the data into multiple categories might help to explore the nature of the interaction effect. Consequently, we recommend application of statistical methods developed for assessment of interactions between continuous biomarkers and treatment instead of arbitrary or data-driven categorization of continuous covariates. [ABSTRACT FROM AUTHOR]