Explainable and Interpretable Machine Learning for Antimicrobial Stewardship: Opportunities and Challenges.

There is growing interest in exploiting the advances in artificial intelligence and machine learning (ML) for improving and monitoring antimicrobial prescriptions in line with antimicrobial stewardship principles. Against this background, the concepts of interpretability and explainability are becoming increasingly essential to understanding how ML algorithms could predict antimicrobial resistance or recommend specific therapeutic agents, to avoid unintended biases related to the "black box" nature of complex models. In this commentary, we review and discuss some relevant topics on the use of ML algorithms for antimicrobial stewardship interventions, highlighting opportunities and challenges, with particular attention paid to interpretability and explainability of employed models. As in other fields of medicine, the exponential growth of artificial intelligence and ML indicates the potential for improving the efficacy of antimicrobial stewardship interventions, at least in part by reducing time-consuming tasks for overwhelmed health care personnel. Improving our knowledge about how complex ML models work could help to achieve crucial advances in promoting the appropriate use of antimicrobials, as well as in preventing antimicrobial resistance selection and dissemination.

Early diagnosis of candidemia with explainable machine learning on automatically extracted laboratory and microbiological data: results of the AUTO-CAND project.

BACKGROUND: Candidemia is associated with a heavy burden of morbidity and mortality in hospitalized patients. The availability of blood culture results could require up to 48-72 h after blood draw; thus, early treatment decisions are made in the absence of a definite diagnosis. METHODS: In this retrospective study, we assessed the performance of different supervised machine learning algorithms for the early differential diagnosis of candidemia and bacteremia in adult patients on a large dataset automatically extracted within the AUTO-CAND project. RESULTS: Overall, 12,483 episodes of candidemia (1275; 10%) or bacteremia (11,208; 90%) were included in the analysis. A random forest classifier achieved the best diagnostic performance for candidemia, with sensitivity 0.98 and specificity 0.65 on the training set (true skill statistic [TSS] = 0.63) and sensitivity 0.74 and specificity 0.57 on the test set (TSS = 0.31). Then, the random classifier was trained in the subgroup of patients with available serum ß-D-glucan (BDG) and procalcitonin (PCT) values by exploiting the feature ranking learned in the entire dataset. Although no statistically significant differences were observed from the performance measures obtained by employing BDG and PCT alone, the performance measures of the classifier that included the features selected in the entire dataset, plus BDG and PCT, were the highest in most cases. CONCLUSIONS: Random forest classifiers trained on large datasets of automatically extracted data have the potential to improve current diagnostic algorithms for candidemia. However, further development through implementation of automatically extracted clinical features may be necessary to achieve crucial improvements.

Pneumocystis jirovecii pneumonia in intensive care units: a multicenter study by ESGCIP and EFISG.

BACKGROUND: Pneumocystis jirovecii pneumonia (PJP) is an opportunistic, life-threatening disease commonly affecting immunocompromised patients. The distribution of predisposing diseases or conditions in critically ill patients admitted to intensive care unit (ICU) and subjected to diagnostic work-up for PJP has seldom been explored. MATERIALS AND METHODS: The primary objective of the study was to describe the characteristics of ICU patients subjected to diagnostic workup for PJP. The secondary objectives were: (i) to assess demographic and clinical variables associated with PJP; (ii) to assess the performance of Pneumocystis PCR on respiratory specimens and serum BDG for the diagnosis of PJP; (iii) to describe 30-day and 90-day mortality in the study population. RESULTS: Overall, 600 patients were included in the study, of whom 115 had presumptive/proven PJP (19.2%). Only 8.8% of ICU patients subjected to diagnostic workup for PJP had HIV infection, whereas hematological malignancy, solid tumor, inflammatory diseases, and solid organ transplants were present in 23.2%, 16.2%, 15.5%, and 10.0% of tested patients, respectively. In multivariable analysis, AIDS (odds ratio [OR] 3.31; 95% confidence interval [CI] 1.13-9.64, p = 0.029), non-Hodgkin lymphoma (OR 3.71; 95% CI 1.23-11.18, p = 0.020), vasculitis (OR 5.95; 95% CI 1.07-33.22, p = 0.042), metastatic solid tumor (OR 4.31; 95% CI 1.76-10.53, p = 0.001), and bilateral ground glass on CT scan (OR 2.19; 95% CI 1.01-4.78, p = 0.048) were associated with PJP, whereas an inverse association was observed for increasing lymphocyte cell count (OR 0.64; 95% CI 0.42-1.00, p = 0.049). For the diagnosis of PJP, higher positive predictive value (PPV) was observed when both respiratory Pneumocystis PCR and serum BDG were positive compared to individual assay positivity (72% for the combination vs. 63% for PCR and 39% for BDG). Cumulative 30-day mortality and 90-day mortality in patients with presumptive/proven PJP were 52% and 67%, respectively. CONCLUSION: PJP in critically ill patients admitted to ICU is nowadays most encountered in non-HIV patients. Serum BDG when used in combination with respiratory Pneumocystis PCR could help improve the certainty of PJP diagnosis.

Mortality in KPC-producing Klebsiella pneumoniae bloodstream infections: a changing landscape.

OBJECTIVES: To assess the impact of carbapenem resistance on mortality in Klebsiella pneumoniae bloodstream infection (BSI) in the era of novel ß-lactam/ß-lactamase inhibitor combinations. MATERIAL AND METHODS: Retrospective study of patients with K. pneumoniae BSI between January and August 2020 in 16 centres (CARBANEW study within the MULTI-SITA project). RESULTS: Overall, 426 patients were included: 107/426 (25%) had carbapenem-resistant K. pneumoniae (CR-Kp) BSI and 319/426 (75%) had carbapenem-susceptible K. pneumoniae (CS-Kp) BSI. Crude cumulative 30 day mortality was 33.8% and 20.7% in patients with, respectively, CR-Kp BSI and CS-Kp BSI (P = 0.027). Carbapenemase production or carbapenemase-encoding genes were detected in 84/98 tested CR-Kp isolates (85.7%), mainly KPC (78/84; 92.9%). Ceftazidime/avibactam was the most frequently used appropriate therapy for CR-Kp BSI (80/107; 74.7%). In multivariable analyses, variables showing an unfavourable association with mortality after correction for multiple testing were age-adjusted Charlson comorbidity index (HR 1.20; 95% CI 1.10-1.31, P < 0.001) and Pitt score (HR 1.33; 95% CI 1.15-1.55, P < 0.001), but not carbapenem resistance (HR 1.28, 95% CI 0.74-2.22, P = 0.410). In a propensity score-matched analysis, there was no difference in mortality between patients appropriately treated with ceftazidime/avibactam for CR-Kp BSI and patients appropriately treated with other agents (mainly meropenem monotherapy or piperacillin/tazobactam monotherapy) for CS-Kp BSI (HR 1.07; 95% CI 0.50-2.29, P = 0.866). CONCLUSIONS: Our results suggest that the increased mortality in CR-Kp BSI compared with CS-Kp BSI is not (or no longer) dependent on the type of therapy in areas where ceftazidime/avibactam-susceptible KPC-producing isolates are the most prevalent type of CR-Kp.

External validation of unsupervised COVID-19 clinical phenotypes and their prognostic impact.

BACKGROUND: Hospitalized patients with coronavirus disease 2019 (COVID-19) can be classified into different clinical phenotypes based on their demographic, clinical, radiology, and laboratory features. We aimed to validate in an external cohort of hospitalized COVID-19 patients the prognostic value of a previously described phenotyping system (FEN-COVID-19) and to assess the reproducibility of phenotypes development as a secondary analysis. METHODS: Patients were classified in phenotypes A, B or C according to the severity of oxygenation impairment, inflammatory response, hemodynamic and laboratory tests according to the FEN-COVID-19 method. RESULTS: Overall, 992 patients were included in the study, and 181 (18%), 757 (76%) and 54 (6%) of them were assigned to the FEN-COVID-19 phenotypes A, B, and C, respectively. An association with mortality was observed for phenotype C vs. A (hazard ratio [HR] 3.10, 95% confidence interval [CI] 1.81-5.30, p < 0.001) and for phenotype C vs. B (HR 2.20, 95% CI 1.50-3.23, p < 0.001). A non-statistically significant trend towards higher mortality was also observed for phenotype B vs. A (HR 1.41; 95% CI 0.92-2.15, p = 0.115). By means of cluster analysis, three different phenotypes were also identified in our cohort, with an overall similar gradient in terms of prognostic impact to that observed when patients were assigned to FEN-COVID-19 phenotypes. CONCLUSIONS: The prognostic impact of FEN-COVID-19 phenotypes was confirmed in our external cohort, although with less difference in mortality between phenotypes A and B than in the original study.

Hospitalized patients with COVID-19 can be classified into different clinical phenotypes based on their demographic, clinical, radiology, and laboratory featuresIn this study, we externally confirmed the prognostic impact of clinical phenotypes previously identified by Gutierrez-Gutierrez and colleagues in a Spanish cohort of hospitalized patients with COVID-19, and the usefulness of their simplified probabilistic model for phenotypes assignmentThis could indirectly support the validity of both phenotype's development and their extrapolation to other hospitals and countries for management decisions during other possible future viral pandemics.

Validation of an Automated System for the Extraction of a Wide Dataset for Clinical Studies Aimed at Improving the Early Diagnosis of Candidemia.

There is increasing interest in assessing whether machine learning (ML) techniques could further improve the early diagnosis of candidemia among patients with a consistent clinical picture. The objective of the present study is to validate the accuracy of a system for the automated extraction from a hospital laboratory software of a large number of features from candidemia and/or bacteremia episodes as the first phase of the AUTO-CAND project. The manual validation was performed on a representative and randomly extracted subset of episodes of candidemia and/or bacteremia. The manual validation of the random extraction of 381 episodes of candidemia and/or bacteremia, with automated organization in structured features of laboratory and microbiological data resulted in ≥99% correct extractions (with confidence interval < ±1%) for all variables. The final automatically extracted dataset consisted of 1338 episodes of candidemia (8%), 14,112 episodes of bacteremia (90%), and 302 episodes of mixed candidemia/bacteremia (2%). The final dataset will serve to assess the performance of different ML models for the early diagnosis of candidemia in the second phase of the AUTO-CAND project.

Increased Mortality Associated with Higher Pre-Dialysis Serum Sodium Variability: Results of the International MONitoring Dialysis Outcome Initiative.

BACKGROUND: Low serum sodium (SNa) is associated with an increased mortality in chronic hemodialysis (HD) patients. Dialysis patients are thought to have an individual pre-dialysis SNa set-point, yet there is evidence for variability of pre-dialysis SNa in individual patient. In this study, we explored the association of several SNa variability metrics with all-cause mortality in a large patient population from the international MONitoring Dialysis Outcomes (MONDO) Initiative. METHODS: All adult incident patients from the MONDO database with more than 5 SNa measurements during the first year on HD were included. All patients were required to survive the first year on HD (defined as the baseline). During the subsequent 2 years of follow-up, all-cause mortality was recorded. The following variability indicators were calculated during baseline: mean SNa and its SD; average real variability (ARV, average the absolute distance of the 2 consecutive SNa measurements), and average directional range (DR, the difference between minimum and maximum values). We used Cox Proportional hazard model with bivariate spline terms to analyze the joint association of SNa and SD, ARV and DR, respectively, with all-cause mortality. While conducting the multivariate Cox regression analyses, patients were stratified into 3 groups of DR (Negative DR: -20≤ DR ≤ -6, Null DR: -6< DR < 6 and Positive DR: 6≤ DR ≤20) with the Null DR as the reference group. RESULTS: We included 20,216 patients in the study. A SNa ≤135 mEq/L was observed to be the strongest predictor of evaluated mortality risk. Higher SNa variability (quantified as SD, ARV, and DR) was also associated with an increased mortality irrespective of SNa levels. When compared with higher SD or ARV, greater DR showed a stronger association with an elevated risk of death. Controlling the Cox Proportional hazard models for additional parameters showed consistent results. CONCLUSION: Higher SNa variability associated with increased all-cause mortality at all levels of SNa. DR of SNa showed the strongest association with mortality and may constitute a Simple and novel prognostic indicator, easily applicable at the bedside.

Microbiota and metabolic diseases.

The microbiota is a complex ecosystem of microorganisms consisting of bacteria, viruses, protozoa, and fungi, living in different districts of the human body, such as the gastro-enteric tube, skin, mouth, respiratory system, and the vagina. Over 70% of the microbiota lives in the gastrointestinal tract in a mutually beneficial relationship with its host. The microbiota plays a major role in many metabolic functions, including modulation of glucose and lipid homeostasis, regulation of satiety, production of energy and vitamins. It exerts a role in the regulation of several biochemical and physiological mechanisms through the production of metabolites and substances. In addition, the microbiota has important anti-carcinogenetic and anti-inflammatory actions. There is growing evidence that any modification in the microbiota composition can lead to several diseases, including metabolic diseases, such as obesity and diabetes, and cardiovascular diseases. This is because alterations in the microbiota composition can cause insulin resistance, inflammation, vascular, and metabolic disorders. The causes of the microbiota alterations and the mechanisms by which microbiota modifications can act on the development of metabolic and cardiovascular diseases have been reported. Current and future preventive and therapeutic strategies to prevent these diseases by an adequate modulation of the microbiota have been also discussed.

Association between Heights of Dialysis Patients and Outcomes: Results from a Retrospective Cohort Study of the International MONitoring Dialysis Outcomes (MONDO) Database Initiative.

BACKGROUND: Tall people have improved metabolic profiles and better cardiovascular outcomes, a relationship inverted in dialysis patients. We investigated the relationship between height and outcomes in incident hemodialysis (HD) patients commencing treatment in an analysis of the international Monitoring Dialysis Outcomes (MONDO) database. METHODS: In this retrospective cohort study, we included incident HD patients commencing treatment between -January 1, 2006 and December 31, 2010 and investigated the association between height and mortality using the MONDO database. A 6-months baseline period preceded 2.5 years of follow-up, during which we recorded patient mortality. Patients were stratified in region-specific deciles of the respective database's population (Asia Pacific, North and South America, and Europe) and we developed Cox-proportional hazard models (additionally adjusted for age, gender, post-dialysis weight, eKt/V, albumin, interdialytic weight gain, phosphorus, and predialysis systolic blood pressure) for each database. RESULTS: We studied 23,353 patients (62 ± 15 years old, 42% female, body mass index 26 ± 6 kg/m2, height 165 ± 10 cm). We found a trend of increasing hazard ratio of death (HR) with increasing height for Asia Pacific, Europe, and South America. In the fully adjusted models, for South America, we found a trend of increasing HR without significance among deciles >5. In Europe, deciles 8-10 had significantly increased HR. No clear trend was found in North America. CONCLUSION: We found an increasing risk of death with increasing height in all regions, except North America. While the reasons remain unclear, further research may be warranted.

Dynamics of the erythropoiesis stimulating agent resistance index in incident hemodiafiltration and high-flux hemodialysis patients.

Hyporesponsiveness to erythropoiesis-stimulating agent therapy in dialysis patients is poorly understood. Some studies report an improvement in the erythropoiesis-stimulating agent resistance index (ERI) with hemodiafiltration (HDF) versus high-flux hemodialysis (HD). We explored ERI dynamics in 38,340 incident HDF and HD patients treated in 22 countries over a 7-year period. Groups were matched by propensity score at baseline (6 months after dialysis initiation). The follow-up period (mean of 1.31 years) was stratified into 1 month intervals with delta analyses performed for key ERI-related parameters. Dialysis modality, time interval, and polycystic kidney disease were included in a linear mixed model with the outcome ERI. Baseline ERI was nonsignificantly higher in HDF versus HD treatment. ERI decreased significantly faster in HDF-treated patients than in HD-treated patients, was decreased in both HD and HDF when patients were treated with intravenous darbepoetin alfa, but only in HDF when treated with intravenous recombinant human erythropoietin (rHuEPO). A clear difference between HD- and HDF-treated patients could only be found for patients with high baseline ERI and assigned to intravenous rHuEPO treatment. A significant advantage in terms of lower ERI for patients treated by HDF was found. Sensitivity analysis limited this advantage for HDF to those patients treated with intravenous rHuEPO (not darbepoetin alfa or subcutaneous rHuEPO) and to patients with a high baseline ERI. Thus, our results allow more accurate planning for future clinical trials addressing anemia management in dialysis patients.

Hemodialysis outcomes in a global sample of children and young adult hemodialysis patients: the PICCOLO MONDO cohort.

BACKGROUND: The aim of this study was to describe the experience of pediatric and young adult hemodialysis (HD) patients from a global cohort. METHODS: The Pediatric Investigation and Close Collaborative Consortium for Ongoing Life Outcomes for MONitoring Dialysis Outcomes (PICCOLO MONDO) study provided de-identified electronic information of 3244 patients, ages 0-30 years from 2000 to 2012 in four regions: Asia, Europe, North America and South America. The study sample was categorized into pediatric (≤18 years old) and young adult (19-30 years old) groups based on the age at dialysis initiation. RESULTS: For those with known end-stage renal disease etiology, glomerular disease was the most common diagnosis in children and young adults. Using Europe as a reference group, North America [odds ratio (OR) 2.69; CI 1.29, 5.63] and South America (OR 4.21; CI 2.32, 7.63) had the greatest mortality among young adults. North America also had higher rates of overweight, obesity, hypertension, cardiovascular disease, hospitalizations and secondary diabetes compared with all other regions. Initial catheter use was greater for North American (86.4% in pediatric patients and 75.2% in young adults) and South America (80.6% in pediatric patients and 75.9% in young adults). Catheter use at 1-year follow-up was most common in North American children (77.3%) and young adults (62.9%). Asia had the lowest rate of catheter use. For both age groups, dialysis adequacy (equilibrated Kt/V) ranged between 1.4 and 1.5. In Asia, patients in both age groups had significantly longer treatment times than in any other region. CONCLUSIONS: The PICCOLO MONDO study has provided unique baseline and 1-year follow-up information on children and young adults receiving HD around the globe. This cohort has brought to light aspects of care in these age groups that warrant further investigation.

Longitudinal Changes in Body Composition in Patients After Initiation of Hemodialysis Therapy: Results From an International Cohort.

OBJECTIVE: In patients with advanced kidney disease, metabolic and nutritional derangements induced by uremia interact and reinforce each other in a deleterious vicious circle. Literature addressing the effect of dialysis initiation on changes in body composition (BC) is limited and contradictory. The aim of this study was to evaluate changes in BC in a large international cohort of incident hemodialysis patients. METHODS: A total of 8,227 incident adult end-stage renal disease patients with BC evaluation within the initial first 6 months of baseline, defined as 6 months after renal replacement therapy initiation, were considered. BC, including fat tissue index (FTI) and lean tissue index (LTI), were evaluated by Body Composition Monitor (BCM, Fresenius Medical Care, Bad Homburg, Germany). Exclusion criteria at baseline were lack of a BCM measurement before or after baseline, body mass index (BMI) < 18.5 kg/m(2), presence of metastatic solid tumors, treatment with a catheter, and prescription of less or more than 3 treatments per week. Maximum follow-up was 2 years. Descriptive analysis was performed comparing current values with the baseline in each interval (delta analysis). Linear mixed models considering the correlation structure of the repeated measurements were used to evaluate factors associated with different trends in FTI and LTI. RESULTS: BMI increased about 0.6 kg/m(2) over 24 months from baseline. This was associated with increase in FTI of about 0.95 kg/m(2) and a decrease in LTI of about 0.4 kg/m(2). Female gender, diabetic status, and low baseline FTI were associated with a significant greater increase of FTI. Age > 67 years, diabetes, male gender, high baseline LTI, and low baseline FTI were associated with a significant greater decrease of LTI. CONCLUSIONS: With the transition to hemodialysis, end-stage renal disease patients presented with distinctive changes in BC. These were mainly associated with gender, older age, presence of diabetes, low baseline FTI, and high baseline LTI. BMI increases did not fully represent the changes in BC.

Seasonal variations in mortality and clinical indicators in international hemodialysis populations from the MONDO registry.

BACKGROUND: Seasonal mortality differences have been reported in US hemodialysis (HD) patients. Here we examine the effect of seasons on mortality, clinical and laboratory parameters on a global scale. METHODS: Databases from the international Monitoring Dialysis Outcomes (MONDO) consortium were queried to identify patients who received in-center HD for at least 1 year. Clinics were stratified by hemisphere and climate zone (tropical or temperate). We recorded mortality and computed averages of pre-dialysis systolic blood pressure (pre-SBP), interdialytic weight gain (IDWG), serum albumin, and log C-reactive protein (CRP). We explored seasonal effects using cosinor analysis and adjusted linear mixed models globally, and after stratification. RESULTS: Data from 87,399 patients were included (northern temperate: 63,671; northern tropical: 7,159; southern temperate: 13,917; southern tropical: 2,652 patients). Globally, mortality was highest in winter. Following stratification, mortality was significantly lower in spring and summer compared to winter in temperate, but not in tropical zones. Globally, pre-SBP and IDWG were lower in summer and spring as compared to winter, although less pronounced in tropical zones. Except for southern temperate zone, serum albumin levels were higher in winter. CRP levels were highest in winter. CONCLUSION: Significant global seasonal variations in mortality, pre-SBP, IDWG, albumin and CRP were observed. Seasonal variations in mortality were most pronounced in temperate climate zones.

Relationship of Neutrophil-to-Lymphocyte Ratio and Serum Albumin Levels with C-Reactive Protein in Hemodialysis Patients: Results from 2 International Cohort Studies.

BACKGROUND/AIM: The neutrophil-to-lymphocyte ratio (NLR), defined as the neutrophil count divided by lymphocyte count, is an inexpensive and readily available parameter, which may serve as a surrogate for inflammation markers, such as C-reactive protein (CRP). The aim of this study was to determine the utility of NLR in the prediction of elevated CRP levels in hemodialysis (HD) patients. METHODS: We analyzed 43,272 HD patients from 2 distinct cohorts within the Monitoring Dialysis Outcomes research collaboration in whom contemporaneous measurements of neutrophil and lymphocyte counts, serum albumin and CRP levels were available. Logistic regression was used to determine the relationship of trichotomized NLR (<2.5, 2.5-5 and >5.0) and albumin levels (<3.1, 3.1-4.0 and >4.0 g/dl) with elevated CRP levels (>10.0, >20.0 and >30.0 mg/l). Congruence of the prediction models was examined by comparing the regression parameters and by cross-validating each regression equation within the other cohort. RESULTS: We found that NLR >5.0 vs. <2.5 (cohort 1: OR 2.3; p < 0.0001 and cohort 2: OR 2.0; p < 0.0001) was associated with CRP levels >10.0 mg/l. Stepwise increase in odds ratio for CRP >10.0 mg/l was observed with the combination of high NLR and low albumin levels (NLR >5.0 and albumin <3.1) (cohort 1: OR 7.6; p < 0.0001 and cohort 2: OR 11.9; p < 0.0001). Cross-validation of the 2 regression models revealed a predictive accuracy of 0.68 and 0.69 in the respective cohorts. CONCLUSION: This study suggests that NLR could serve as a potential surrogate marker for CRP. Our results may add to diagnostic abilities in settings where CRP is not measured routinely in HD patients. NLR is easy to integrate into daily practice and may be used as a marker of systemic inflammation.

Plasma clusterin as a candidate pre-diagnosis marker of colorectal cancer risk in the Florence cohort of the European Prospective Investigation into Cancer and Nutrition: a pilot study.

BACKGROUND: Colorectal cancer is one of the major causes of cancer mortality world-wide. Prevention would improve if at-risk subjects could be identified. The aim of this study was to characterise plasma protein biomarkers associated with the risk of colorectal cancer in samples collected prospectively, before the disease diagnosis. METHODS: After an exploratory study on the comprehensive plasma proteome analysis by liquid chromatography-tandem mass spectrometry from ten colorectal cancer cases enrolled at diagnosis, and ten matched controls (Phase 1), a similar preliminary study was performed on prospective plasma samples from ten colorectal cancer cases, enrolled years before disease development, and ten matched controls identified in a nested case-control study within the Florence cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC) study (Phase 2); in Phase 3 the validation of the candidate biomarkers by targeted proteomics on 48 colorectal cancer cases and 48 matched controls from the Florence-EPIC cohort, and the evaluation of the disease risk were performed. RESULTS: Systems biology tools indicated that both in the Phase 1 and Phase 2 studies circulating protein levels differing in cases more than 1.5 times from controls, were involved in inflammation and/or immune response. Eight proteins including apolipoprotein C-II, complement C4-B, complement component C9, clusterin, alpha-2-HS-glycoprotein, mannan-binding lectin serine-protease, mannose-binding protein C, and N-acetylmuramoyl-L-alanine amidase were selected as promising candidate biomarkers. Targeted proteomics of the selected proteins in the EPIC samples showed significantly higher clusterin levels in cases than controls, but only in men (mean ± SD, 1.98 ± 0.46 and 1.61 ± 0.43 nmol/mL respectively, Mann-Whitney U, two-tailed P = 0.0173). The remaining proteins were unchanged. Using multivariate logistic models a significant positive association emerged for clusterin, with an 80% increase in the colorectal cancer risk with protein's unit increase, but only in men. CONCLUSIONS: The results show that plasma proteins can be altered years before colorectal cancer detection. The high circulating clusterin in pre-diagnostic samples suggests this biomarker can improve the identification of people at risk of colorectal cancer and might help in designing preventive interventions.

Body composition and survival in dialysis patients: results from an international cohort study.

BACKGROUND AND OBJECTIVES: High body mass index appears protective in hemodialysis patients, but uncertainty prevails regarding which components of body composition, fat or lean body mass, are primarily associated with survival. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: Data between April 2006 and December 2012 were extracted from the Fresenius Medical Care Europe subset of the international MONitoring Dialysis Outcomes initiative. Fresenius Medical Care Europe archives a unique repository of predialysis body composition measurements determined by multifrequency bioimpedance (BCM Body Composition Monitor). The BCM Body Composition Monitor reports lean tissue indices (LTIs) and fat tissue indices (FTIs), which are the respective tissue masses normalized to height squared, relative to an age- and sex-matched healthy population. The relationship between LTI and FTI and all-cause mortality was studied by Kaplan-Meier analysis, multivariate Cox regression, and smoothing spline ANOVA logistic regression. RESULTS: In 37,345 hemodialysis patients, median (25th-75th percentile) LTI and FTI were 12.2 (10.3-14.5) and 9.8 (6.6-12.4) kg/m(2), respectively. Median (25th-75th percentile) follow-up time was 266 (132-379) days; 3458 (9.2%) patients died during follow-up. Mortality was lowest with both LTI and FTI in the 10th-90th percentile (reference group) and significantly higher at the lower LTI and FTI extreme (hazard ratio [HR], 3.37; 95% confidence interval [95% CI], 2.94 to 3.87; P<0.001). Survival was best with LTI between 15 and 20 kg/m(2) and FTI between 4 and 15 kg/m(2) (probability of death during follow-up: <5%). When taking the relation between both compartments into account, the interaction was significant (P=0.01). Higher FTI appeared protective in patients with low LTI (HR, 3.37; 95% CI, 2.94 to 3.87; P<0.001 at low LTI-low FTI, decreasing to HR, 1.79; 95% CI, 1.47 to 2.17; P<0.001 at low LTI-high FTI). CONCLUSIONS: This large international study indicates best survival in patients with both LTI and FTI in the 10th-90th percentiles of a healthy population. In analyses of body composition, both lean tissue and fat tissue compartments and also their relationship should be considered.

Nutritional Markers and Body Composition in Hemodialysis Patients.

The aims of this study were to analyse body composition, to detect the presence of undernutrition, and to establish a relationship between undernutrition and the biological markers routinely used as indicators of nutritional status in hemodialysis (HD) patients (pts). We used a body composition monitor (BCM) that expresses body weight in terms of lean tissue mass (LTM) and fat tissue mass (FTM) independent of hydration status. From nine HD units, 934 pts were included. Undernutrition was defined as having a lean tissue index (LTI = LTM/height(2)) below the 10th percentile of a reference population. Biochemical markers and parameters delivered by BCM were used to compare low LTI and normal LTI groups. Undernutrition prevalence was 58.8% of the population studied. Low LTI pts were older, were significantly more frequently overhydrated, and had been on HD for a longer period of time than the normal LTI group. FTI (FTI = FTM/ height(2)) was significantly higher in low LTI pts and increased according to BMI. LTI was not influenced by different BMI levels. Albumin and C-reactive protein correlated inversely (r = -0.28). However neither of them was statistically different when considering undernourished and normal LTI pts. Our BCM study was able to show a high prevalence of undernutrition, as expressed by low LTI. In our study, BMI and other common markers, such as albumin, failed to predict malnutrition as determined by BCM.

Interdialytic weight gain, systolic blood pressure, serum albumin, and C-reactive protein levels change in chronic dialysis patients prior to death.

Reports from a United States cohort of chronic hemodialysis patients suggested that weight loss, a decline in pre-dialysis systolic blood pressure, and decreased serum albumin may precede death. However, no comparative studies have been reported in such patients from other countries. Here we analyzed dynamic changes in these parameters in hemodialysis patients and included 3593 individuals from 5 Asian countries; 35,146 from 18 European countries; 8649 from Argentina; and 4742 from the United States. In surviving prevalent patients, these variables appeared to have notably different dynamics than in patients who died. While in all populations the interdialytic weight gain, systolic blood pressure, and serum albumin levels were stable in surviving patients, these indicators declined starting more than a year ahead in those who died with the dynamics similar irrespective of gender and geographic region. In European patients, C-reactive protein levels were available on a routine basis and indicated that levels of this acute-phase protein were low and stable in surviving patients but rose sharply before death. Thus, relevant fundamental biological processes start many months before death in the majority of chronic hemodialysis patients. Longitudinal monitoring of these dynamics may help to identify patients at risk and aid the development of an alert system to initiate timely interventions to improve outcomes.

The MONitoring Dialysis Outcomes (MONDO) initiative.

BACKGROUND: Systematic collection and analysis of global hemodialysis patient data may help to improve patient outcomes. METHODS: The MONitoring Dialysis Outcomes (MONDO) initiative comprises data from eight dialysis providers worldwide. Data are combined into one repository. Extensive procedures are employed to merge data across countries and providers. RESULTS: The MONDO database comprises longitudinal data of currently 128,000 hemodialysis patients from 26 countries on five continents. Here we report data from 62,345 incident hemodialysis patients. We found lower catheter rates in South-East Asia and Australia, lower hemoglobin levels in South-East Asia, and a higher prevalence of diabetes in North America. Longitudinal analyses suggest that there is a decline in interdialytic weight gain and serum phosphorus and an increasing neutrophil-to-lymphocyte ratio before death in all regions studied. CONCLUSIONS: While organizationally lean and low-cost, MONDO is the largest global dialysis database initiative to date, with a particular focus on high longitudinal data density and geographical diversity.

Monitoring dialysis outcomes across the world--the MONDO Global Database Consortium.

BACKGROUND/AIMS: Dialysis providers frequently collect detailed longitudinal and standardized patient data, providing valuable registries of routine care. However, even large organizations are restricted to certain regions, limiting their ability to separate effects of local practice from the pathophysiology shared by most dialysis patients. To overcome this limitation, the MONDO (MONitoring Dialysis Outcomes) research consortium has created a platform for the joint analysis of data from almost 200,000 dialysis patients worldwide. METHODS: We examined design and operation of MONDO as well as its methodology with respect to patient inclusion, descriptive data and other study parameters. RESULTS: MONDO partners contribute primary databases of anonymized patient data and collaboratively analyze populations across national and regional boundaries. To that end, datasets from different electronic health record systems are converted into a uniform structure. Patients are enrolled without systematic exclusions into open cohorts representing the diversity of patients. A large number of patient level treatment and outcome data is recorded frequently and can be analyzed with little delay. Detailed variable definitions are used to determine if a parameter can be studied in a subset or all databases. CONCLUSION: MONDO has created a large repository of validated dialysis data, expanding the opportunities for outcome studies in dialysis patients. The density of longitudinal information facilitates in particular trend analysis. Limitations include the paucity of uniform definitions and standards regarding descriptive information (e.g. comorbidities), which limits the identification of patient subsets. Through its global outreach, depth, breadth and size, MONDO advances the observational study of dialysis patients and care.