Machine learning derived serum creatinine trajectories in acute kidney injury in critically ill patients with sepsis.

BACKGROUND: Current classification for acute kidney injury (AKI) in critically ill patients with sepsis relies only on its severity-measured by maximum creatinine which overlooks inherent complexities and longitudinal evaluation of this heterogenous syndrome. The role of classification of AKI based on early creatinine trajectories is unclear. METHODS: This retrospective study identified patients with Sepsis-3 who developed AKI within 48-h of intensive care unit admission using Medical Information Mart for Intensive Care-IV database. We used latent class mixed modelling to identify early creatinine trajectory-based classes of AKI in critically ill patients with sepsis. Our primary outcome was development of acute kidney disease (AKD). Secondary outcomes were composite of AKD or all-cause in-hospital mortality by day 7, and AKD or all-cause in-hospital mortality by hospital discharge. We used multivariable regression to assess impact of creatinine trajectory-based classification on outcomes, and eICU database for external validation. RESULTS: Among 4197 patients with AKI in critically ill patients with sepsis, we identified eight creatinine trajectory-based classes with distinct characteristics. Compared to the class with transient AKI, the class that showed severe AKI with mild improvement but persistence had highest adjusted risks for developing AKD (OR 5.16; 95% CI 2.87-9.24) and composite 7-day outcome (HR 4.51; 95% CI 2.69-7.56). The class that demonstrated late mild AKI with persistence and worsening had highest risks for developing composite hospital discharge outcome (HR 2.04; 95% CI 1.41-2.94). These associations were similar on external validation. CONCLUSIONS: These 8 classes of AKI in critically ill patients with sepsis, stratified by early creatinine trajectories, were good predictors for key outcomes in patients with AKI in critically ill patients with sepsis independent of their AKI staging.

Constructing synthetic datasets with generative artificial intelligence to train large language models to classify acute renal failure from clinical notes.

OBJECTIVES: To compare performances of a classifier that leverages language models when trained on synthetic versus authentic clinical notes. MATERIALS AND METHODS: A classifier using language models was developed to identify acute renal failure. Four types of training data were compared: (1) notes from MIMIC-III; and (2, 3, and 4) synthetic notes generated by ChatGPT of varied text lengths of 15 (GPT-15 sentences), 30 (GPT-30 sentences), and 45 (GPT-45 sentences) sentences, respectively. The area under the receiver operating characteristics curve (AUC) was calculated from a test set from MIMIC-III. RESULTS: With RoBERTa, the AUCs were 0.84, 0.80, 0.84, and 0.76 for the MIMIC-III, GPT-15, GPT-30- and GPT-45 sentences training sets, respectively. DISCUSSION: Training language models to detect acute renal failure from clinical notes resulted in similar performances when using synthetic versus authentic training data. CONCLUSION: The use of training data derived from protected health information may not be needed.

Utilizing reclassification to explore characteristics and prognosis of KDIGOSCr AKI subgroups: a retrospective analysis of a multicenter prospective cohort study.

BACKGROUND: Acute kidney injury (AKI) is widespread in the intensive care unit (ICU) and affects patient prognosis. According to Kidney Disease: Improving Global Outcomes (KDIGO) guidelines, the absolute and relative increases of serum creatinine (Scr) are classified into the same stage. Whether the prognosis of the two types of patients is similar in the ICU remains unclear. METHODS: According to the absolute and relative increase of Scr, AKI stage 1 and stage 3 patients were divided into stage 1a and 1b, stage 3a and 3b groups, respectively. Their demographics, laboratory results, clinical characteristics, and outcomes were analyzed retrospectively. RESULTS: Of the 345 eligible cases, we analyzed stage 1 because stage 3a group had only one patient. Using 53 or 61.88 µmol/L as the reference Scr (Scrref), no significant differences were observed in ICU mortality (P53=0.076, P61.88=0.070) or renal replacement therapy (RRT) ratio, (P53=0.356, P61.88=0.471) between stage 1a and 1b, but stage 1b had longer ICU length of stay (LOS) than stage 1a (P53<0.001, P61.88=0.032). In the Kaplan-Meier survival analysis, no differences were observed in ICU mortality between stage 1a and 1b (P53=0.378, P61.88=0.255). In a multivariate analysis, respiratory failure [HR = 4.462 (95% CI 1.144-17.401), p = 0.031] and vasoactive drug therapy [HR = 4.023 (95% CI 1.584-10.216), p = 0.003] were found to be independently associated with increased risk of death. CONCLUSION: ICU LOS benefit was more prominent in KDIGOSCr AKI stage 1a patients than in stage 1 b. Further prospective studies with a larger sample size are necessary to confirm the effectiveness of reclassification.

Relationship between acute kidney injury and mortality in poisoning - a systematic review and metanalysis.

RATIONALE: Three consensus classifications of acute kidney injury have been published. These are RIFLE (Risk, Injury, Failure, Loss of kidney function, and End-stage kidney disease published by the Acute Dialysis Quality Initiative workgroup), AKIN (published by the Acute Kidney Injury Network) and KDIGO (published by the Kidney Disease Improving Global Outcome workgroup). Acute kidney injury has been reported consistently as associated with worsened outcomes. However, toxicant-related acute kidney injury has been excluded from the studies used to validate the classifications of acute kidney injury. OBJECTIVE: To study whether poisoned patients who develop acute kidney injury, as defined by consensus definitions/classifications, have higher mortality compared to those who did not. METHODS: Databases were searched from 2004 to 2019 using the following keywords (KDIGO OR "Kidney Disease: Improving Global Outcomes" OR "Kidney Disease Improving Global Outcomes" OR AKIN OR "AKI network" OR "Acute kidney Injury Network" OR ADQI OR RIFLE OR "Acute dialysis quality initiative") AND (intoxication OR poisoning OR overdose OR ingestion) AND (AKI OR kidney OR renal OR ARF). If data were available, we used a random-effects meta-analysis model and Fisher's exact test to compare mortality in patients according to kidney function definitions (acute kidney injury vs not) and stages (stages vs no acute kidney injury), respectively. If data were available, we assessed the correlation between mortality and renal function (no acute kidney injury, risk/stage 1, injury/stage 2 and failure/stage 3) using the Spearman correlation. If available, we collected the results of statistical analyses in studies that have used acute kidney injury to predict mortality. RESULTS: Study selection. Thirty-three relevant studies were found, 22/33 retrospective studies (67%) and 11/33 prospective studies (33%). Paraquat was the most frequent toxicant involved (13/33, 39%). We found a disparity between studies regarding the timeframe during which mortality was assessed, the temporality of the renal function considered to predict mortality (initial/worst) and the criteria used to define/grade acute kidney injury across studies. Univariate association between acute kidney injury definitions/stages and mortality. Consensus definitions/staging of acute kidney injury were associated with higher mortality, using univariate analyses, in twenty-eight (RIFLE = 7; AKIN = 12; KDIGO = 9) studies included but not in five (AKIN = 4, KDIGO = 1). When available data were pooled, RIFLE (5 studies), AKIN (16 studies) and KDIGO definitions (8 studies) of acute kidney injury were associated with a higher mortality (Log unadjusted Odds ratios [95%-confidence interval], 2.60 [2.23; 2.97], 2.02 [1.48; 2,52] and 3.22 [2,65; 3.78], respectively). However, we found high heterogeneity (I2=54,7%) and publication bias among studies using AKIN. In ten studies with available data, the correlation between renal function (no acute kidney injury, risk/stage 1, injury/stage 2, failure/stage 3) and mortality was significant in 5 studies (RIFLE = 2; AKIN = 3), but not in five studies (RIFLE = 1; AKIN = 3; KDIGO = 1).Multivariate association between acute kidney injury definitions/stages and mortality. The definitions of acute kidney injury were associated with higher mortality in two studies (RIFLE = 2), but not in four studies (AKIN = 1 and KDIGO = 3. The stages of acute kidney injury (including one or more stages) were associated with higher mortality in four (RIFLE = 1, AKIN = 1 and KDIGO = 2). CONCLUSIONS: All three consensus definitions/classifications were associated independently with increased mortality in poisoning but with disparity between studies reporting acute kidney injury.

Definition of acute kidney injury impacts prevalence and prognosis in ACS patients undergoing coronary angiography.

BACKGROUND: Development of acute kidney injury (AKI) in invasively managed patients with acute coronary syndrome (ACS) is associated with a markedly increased mortality risk. Different definitions of AKI are in use, leading to varying prevalence and outcome measurements. The aim of the present study is to analyze an ACS population undergoing coronary angiography for differences in AKI prevalence and outcome using four established AKI definitions. METHODS: 944 patients (30% female) were enrolled in a prospective registry between 2003 and 2005 with 6-month all-cause mortality as outcome measure. Four established AKI definitions were used: an increase in serum creatinine (sCR) ≥ 1.5 fold, ≥ 0.3 mg/dl, and ≥ 0.5 mg/dl and a decrease in eGFR > 25% from baseline (AKIN 1, AKIN 2, CIN, and RIFLE definition groups, respectively). RESULTS: AKI rates varied widely between the different groups. Using the CIN definition, AKI frequency was lowest (4.4%), whereas it was highest if the RIFLE definition was applied (13.2%). AKIN 2 displayed a twofold higher AKI prevalence compared with AKIN 1 (10.2% vs. 5.3% (p < 0.001)). AKI was a strong risk factor for mid-term mortality, with distinctive variability between the definitions. The lowest mortality risk was found in the RIFLE group (HR 6.0; 95% CI 3.7-10.0; p < 0.001), whereas CIN revealed the highest risk (HR 16.7; 95% CI 9.9-28.1; p < 0.001). CONCLUSION: Prevalence and outcome in ACS patients varied considerably depending on the AKI definition applied. To define patients with highest renal function-associated mortality risk, use of the CIN definition seems to have the highest prognostic relevance.

Conceptual advances and evolving terminology in acute kidney disease.

Over the past decade, new insights into epidemiology, pathophysiology and biomarkers have modified our understanding of acute kidney dysfunction and damage, and their association with subsequent chronic kidney disease. The concept of acute kidney injury (AKI), which has relied on established but nonetheless flawed biomarkers of solute clearance (serum creatinine levels and urinary output), has been challenged by the identification of novel biomarkers of tubular stress and/or damage. The expression of some of these novel biomarkers precedes changes in conventional biomarkers or can increase their predictive power, and might therefore enhance the clinical accuracy of the definition of AKI. In addition, the need to consider AKI recurrence, duration and progression to chronic kidney disease within the clinical and epidemiological framework of AKI led to the emergence of the concept of acute kidney disease. New definitions of acute syndromes of kidney impairment and injury are needed.

The Impact of Erythropoietin on Short- and Long-Term Kidney-Related Outcomes in Neonates of Extremely Low Gestational Age. Results of a Multicenter, Double-Blind, Placebo-Controlled Randomized Clinical Trial.

OBJECTIVE: To evaluate whether extremely low gestational age neonates (ELGANs) randomized to erythropoietin have better or worse kidney-related outcomes during hospitalization and at 22-26 months of corrected gestational age (cGA) compared with those randomized to placebo. STUDY DESIGN: We performed an ancillary study to a multicenter double-blind, placebo-controlled randomized clinical trial of erythropoietin in ELGANs. RESULTS: The prevalence of severe (stage 2 or 3) acute kidney injury (AKI) was 18.2%. We did not find a statistically significant difference between those randomized to erythropoietin vs placebo for in-hospital primary (severe AKI) or secondary outcomes (any AKI and serum creatinine/cystatin C values at days 0, 7, 9, and 14). At 22-26 months of cGA, 16% of the cohort had an estimated glomerular filtration rate (eGFR) <90 mL/min/1.73 m2, 35.8% had urine albumin/creatinine ratio >30 mg/g, 23% had a systolic blood pressure (SBP) >95th percentile for age, and 40% had a diastolic blood pressure (DBP) >95th percentile for age. SBP >90th percentile occurred less often among recipients of erythropoietin (P < .04). This association remained even after controlling for gestational age, site, and sibship (aOR 0.6; 95% CI 0.39-0.92). We did not find statistically significant differences between treatment groups in eGFR, albumin/creatinine ratio, rates of SBP >95th percentile, or DBP >90th or >95th percentiles at the 2 year follow-up visit. CONCLUSIONS: ELGANs have high rates of in-hospital AKI and kidney-related problems at 22-26 months of cGA. Recombinant erythropoietin may protect ELGANs against long-term elevated SBP but does not appear to protect from AKI, low eGFR, albuminuria, or elevated DBP at 22-26 months of cGA.

Dividing Stage 1 Acute Kidney Injury After Cardiac or Thoracic Aortic Surgery.

BACKGROUND: Although the Kidney Disease: Improving Global Outcomes (KDIGO) criteria are used to define acute kidney injury, the criteria have limitations for including 2 different serum creatinine criteria in stage 1. We hypothesized that there would be differences in clinical outcomes between the 2 subgroups of stage 1 acute kidney injury in patients undergoing cardiac or thoracic aortic surgery. METHODS: We reviewed 2510 cases. Patients with KDIGO stage 1 were divided into 2 subgroups (stage 1a: 0.3 mg/dL or greater of absolute increase in serum creatinine, n = 376; and stage 1b: 50% or greater relative increase, n = 365). Propensity score analysis was performed between stage 1a and 1b groups, yielding 240 pairs. We compared the length of hospital stay, the incidence of cardiovascular complications, 5-year all-cause mortality between these subgroups. Overall survival was compared between the subgroups after propensity score matching. We performed sensitivity analysis for Acute Kidney Injury Network (AKIN) criteria. RESULTS: Length of hospital stay and 5-year all-cause mortality were worse in patients with KDIGO stage 1b compared with stage 1a. Five-year patient survival was significantly worse in patients with stage 1b compared with stage 1a after matching (log rank test, P = .002). We found similar results regarding AKIN criteria. Subgroup analysis showed that the significant difference in survival existed only when baseline serum creatinine was 0.8 mg/dL or greater. CONCLUSIONS: The KDIGO or AKIN criteria for stage 1 acute kidney injury could be further divided into 2 substages with different severity of clinical outcomes. These modified criteria could give additional prognostic information in patients undergoing cardiac or thoracic aortic surgery.

One concept does not fit all: the immune system in different forms of acute kidney injury.

Renal and immune systems maintain body homoeostasis during physiological fluctuations and following tissue injury. The immune system plays a central role during acute kidney injury (AKI), adapting evolutional systems programmed for host defence and minimizing unnecessary collateral damage. Indeed, depending upon the disease context, the impact of the immune system upon the manifestations and consequences of AKI can be quite different. Here we provide an overview of the known and unknown involvement of the immune system within the wide range of different forms of AKI, to oppose oversimplification and to endorse deeper insights into the pathogenesis of the different diseases causing kidney injury. This approach may help to overcome some of the current hurdles in translational AKI research and the development of specific treatments for the different diseases, all presenting with an acute increase in serum creatinine or decline in urinary output. One concept does not fit all.

Sub-Phenotypes of Acute Kidney Injury: Do We Have Progress for Personalizing Care?

Acute kidney injury (AKI) is the most common form of organ dysfunction occurring in patients admitted to the intensive care unit and contributes significantly to poor long-term outcomes. Despite this public health impact, no effective pharmacotherapy exists for AKI. One reason may be that heterogeneity is present within AKI as currently defined, thereby concealing unique pathophysiologic processes specific to certain AKI populations. Supporting this notion, we and others have shown that diversity within the AKI clinical syndrome exists, and the "one-size-fits-all" approach by current diagnostic guidelines may not be ideal. A "precision medicine" approach that exploits an individual's genetic, biologic, and clinical characteristics to identify AKI sub-phenotypes may overcome such limitations. Identification of AKI sub-phenotypes may address a critical unmet clinical need in AKI by (1) improving risk prognostication, (2) identifying novel pathophysiology, and (3) informing a patient's likelihood of responding to current therapeutics or establishing new therapeutic targets to prevent and treat AKI. This review discusses the current state of phenotyping AKI and future directions.

Utilization of Deep Learning for Subphenotype Identification in Sepsis-Associated Acute Kidney Injury.

BACKGROUND AND OBJECTIVES: Sepsis-associated AKI is a heterogeneous clinical entity. We aimed to agnostically identify sepsis-associated AKI subphenotypes using deep learning on routinely collected data in electronic health records. DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS: We used the Medical Information Mart for Intensive Care III database, which consists of electronic health record data from intensive care units in a tertiary care hospital in the United States. We included patients ≥18 years with sepsis who developed AKI within 48 hours of intensive care unit admission. We then used deep learning to utilize all available vital signs, laboratory measurements, and comorbidities to identify subphenotypes. Outcomes were mortality 28 days after AKI and dialysis requirement. RESULTS: We identified 4001 patients with sepsis-associated AKI. We utilized 2546 combined features for K-means clustering, identifying three subphenotypes. Subphenotype 1 had 1443 patients, and subphenotype 2 had 1898 patients, whereas subphenotype 3 had 660 patients. Subphenotype 1 had the lowest proportion of liver disease and lowest Simplified Acute Physiology Score II scores compared with subphenotypes 2 and 3. The proportions of patients with CKD were similar between subphenotypes 1 and 3 (15%) but highest in subphenotype 2 (21%). Subphenotype 1 had lower median bilirubin levels, aspartate aminotransferase, and alanine aminotransferase compared with subphenotypes 2 and 3. Patients in subphenotype 1 also had lower median lactate, lactate dehydrogenase, and white blood cell count than patients in subphenotypes 2 and 3. Subphenotype 1 also had lower creatinine and BUN than subphenotypes 2 and 3. Dialysis requirement was lowest in subphenotype 1 (4% versus 7% [subphenotype 2] versus 26% [subphenotype 3]). The mortality 28 days after AKI was lowest in subphenotype 1 (23% versus 35% [subphenotype 2] versus 49% [subphenotype 3]). After adjustment, the adjusted odds ratio for mortality for subphenotype 3, with subphenotype 1 as a reference, was 1.9 (95% confidence interval, 1.5 to 2.4). CONCLUSIONS: Utilizing routinely collected laboratory variables, vital signs, and comorbidities, we were able to identify three distinct subphenotypes of sepsis-associated AKI with differing outcomes.

Genetic variation implicates plasma angiopoietin-2 in the development of acute kidney injury sub-phenotypes.

BACKGROUND: We previously identified two acute kidney injury (AKI) sub-phenotypes (AKI-SP1 and AKI-SP2) with different risk of poor clinical outcomes and response to vasopressor therapy. Plasma biomarkers of endothelial dysfunction (tumor necrosis factor receptor-1, angiopoietin-1 and 2) differentiated the AKI sub-phenotypes. However, it is unknown whether these biomarkers are simply markers or causal mediators in the development of AKI sub-phenotypes. METHODS: We tested for associations between single-nucleotide polymorphisms within the Angiopoietin-1, Angiopoietin-2, and Tumor Necrosis Factor Receptor 1A genes and AKI- SP2 in 421 critically ill subjects of European ancestry. Top performing single-nucleotide polymorphisms (FDR < 0.05) were tested for cis-biomarker expression and whether genetic risk for AKI-SP2 is mediated through circulating biomarkers. We also completed in vitro studies using human kidney microvascular endothelial cells. Finally, we calculated the renal clearance of plasma biomarkers using 20 different timed urine collections. RESULTS: A genetic variant, rs2920656C > T, near ANGPT2 was associated with reduced risk of AKI-SP2 (odds ratio, 0.45; 95% CI, 0.31-0.66; adjusted FDR = 0.003) and decreased plasma angiopoietin-2 (p = 0.002). Causal inference analysis showed that for each minor allele (T) the risk of developing AKI-SP2 decreases by 16%. Plasma angiopoietin-2 mediated 41.5% of the rs2920656 related risk for AKI-SP2. Human kidney microvascular endothelial cells carrying the T allele of rs2920656 produced numerically lower levels of angiopoietin-2 although this was not statistically significant (p = 0.07). Finally, analyses demonstrated that angiopoietin-2 is minimally renally cleared in critically ill subjects. CONCLUSION: Genetic mediation analysis provides supportive evidence that angiopoietin-2 plays a causal role in risk for AKI-SP2.

Subphenotypes in critical care: translation into clinical practice.

Despite progress in the supportive care available for critically ill patients, few advances have been made in the search for effective disease-modifying therapeutic options. The fact that many trials in critical care medicine have not identified a treatment benefit is probably due, in part, to the underlying heterogeneity of critical care syndromes. Numerous approaches have been proposed to divide populations of critically ill patients into more meaningful subgroups (subphenotypes), some of which might be more useful than others. Subclassification systems driven by clinical features and biomarkers have been proposed for acute respiratory distress syndrome, sepsis, acute kidney injury, and pancreatitis. Identifying the systems that are most useful and biologically meaningful could lead to a better understanding of the pathophysiology of critical care syndromes and the discovery of new treatment targets, and allow recruitment in future therapeutic trials to focus on predicted responders. This Review discusses proposed subphenotypes of critical illness syndromes and highlights the issues that will need to be addressed to translate subphenotypes into clinical practice.

Different applications of the KDIGO criteria for AKI lead to different incidences in critically ill patients: a post hoc analysis from the prospective observational SICS-II study.

BACKGROUND: Acute kidney injury (AKI) is a frequent and clinically relevant problem in critically ill patients. Various randomized controlled trials (RCT) have attempted to assess potentially beneficial treatments for AKI. Different approaches to applying the Kidney Disease Improving Global Outcomes (KDIGO) criteria for AKI make a comparison of studies difficult. The objective of this study was to assess how different approaches may impact estimates of AKI incidence and whether the association between AKI and 90-day mortality varied by the approach used. METHODS: Consecutive acutely admitted adult intensive care patients were included in a prospective observational study. AKI was determined following the KDIGO criteria during the first 7 days of ICU admission. In this post hoc analysis, we assessed whether AKI incidence differed when applying the KDIGO criteria in 30 different possible methods, varying in (A) serum creatinine (sCr), (B) urine output (UO), and (C) the method of combining these two into an outcome, e.g., severe AKI. We assessed point estimates and 95% confidence intervals for each incidence. Univariable regression was used to assess the associations between AKI and 90-day mortality. RESULTS: A total of 1010 patients were included. Baseline creatinine was available in 449 (44%) patients. The incidence of any AKI ranged from 28% (95%CI 25-31%) to 75% (95%CI 72-77%) depending on the approach used. Methods to estimate missing baseline sCr caused a variation in AKI incidence up to 15%. Different methods of handling UO caused a variation of up to 35%. At 90 days, 263 patients (26%) had died, and all 30 variations were associated with 90-day mortality. CONCLUSIONS: In this cohort of critically ill patients, AKI incidence varied from 28 to 75%, depending on the method used of applying the KDIGO criteria. A tighter adherence to KDIGO definitions is warranted to decrease the heterogeneity of AKI and increase the comparability of future studies.

Delayed versus early initiation of renal replacement therapy for severe acute kidney injury: a systematic review and individual patient data meta-analysis of randomised clinical trials.

BACKGROUND: The timing of renal replacement therapy (RRT) for severe acute kidney injury is highly debated when no life-threatening complications are present. We assessed whether a strategy of delayed versus early RRT initiation affects 28-day survival in critically ill adults with severe acute kidney injury. METHODS: In this systematic review and individual patient data meta-analysis, we searched MEDLINE (via PubMed), Embase, and the Cochrane Central Register of Controlled Trials for randomised trials published from April 1, 2008, to Dec 20, 2019, that compared delayed and early RRT initiation strategies in patients with severe acute kidney injury. Trials were eligible for inclusion if they included critically ill patients aged 18 years or older with acute kidney injury (defined as a Kidney Disease: Improving Global Outcomes [KDIGO] acute kidney injury stage 2 or 3, or, where KDIGO was unavailable, a renal Sequential Organ Failure Assessment score of 3 or higher). We contacted the principal investigator of each eligible trial to request individual patient data. From the included trials, any patients without acute kidney injury or who were not randomly allocated were not included in the individual patient data meta-analysis. The primary outcome was all-cause mortality at day 28 after randomisation. This study is registered with PROSPERO (CRD42019125025). FINDINGS: Among the 1031 studies identified, one study that met the eligibility criteria was excluded because the recruitment period was not recent enough, and ten (including 2143 patients) were included in the analysis. Individual patient data were available for nine studies (2083 patients), from which 1879 patients had severe acute kidney injury and were randomly allocated: 946 (50%) to the delayed RRT group and 933 (50%) to the early RRT group. 390 (42%) of 929 patients allocated to the delayed RRT group and who had available data did not receive RRT. The proportion of patients who died by day 28 did not significantly differ between the delayed RRT group (366 [44%] of 837) and the early RRT group (355 [43%] of 827; risk ratio 1·01 [95% CI 0·91 to 1·13], p=0·80), corresponding to an overall risk difference of 0·01 (95% CI -0·04 to 0·06). There was no heterogeneity across studies (I2=0%; τ2=0), and most studies had a low risk of bias. INTERPRETATION: The timing of RRT initiation does not affect survival in critically ill patients with severe acute kidney injury in the absence of urgent indications for RRT. Delaying RRT initiation, with close patient monitoring, might lead to a reduced use of RRT, thereby saving health resources. FUNDING: None.

Long-term predictive value of acute kidney injury classification in diffuse proliferative lupus nephritis with acute kidney injury.

BACKGROUND: The long-term predictive ability of acute kidney injury (AKI) classification based on "Kidney Disease: Improving Global Outcomes"(KDIGO) AKI diagnosis criteria has not been clinically validated in diffuse proliferative lupus nephritis (DPLN) patients with AKI. Our objective was to assess the long-term predictive value of KDIGO AKI classification in DPLN patients with AKI. METHODS: Retrospective cohort study was conducted by reviewing medical records of biopsy-proven DPLN patients with AKI from the First Affiliated Hospital of Wenzhou Medical University between Jan 1, 2000 and Dec 31, 2014. Multivariate Cox regression and survival analysis were performed. RESULTS: One hundred sixty-seven DPLN patients were enrolled,82(49%) patients were normal renal function (No AKI), 40(24%) patients entered AKI-1 stage (AKI-1), 26(16%) patients entered AKI-2 stage (AKI-2) and 19(16%) patients entered AKI-3 stage (AKI-3). The mean follow-up of all patients was 5.1 ± 3.8 years. The patient survival without ESRD of all patients was 86% at 5 years and 79% at 10 years. The patient survival rate without ESRD at 10 yr was 94.5% for No AKI patients, 81.8% for AKI-1 patients, 44.9% for AKI-2 patients and 14.6% for AKI-3 patients. The area under the ROC curve for KDIGO AKI classification to predict the primary end point was 0.83 (95% CI: 0.73-0.93) (P < 0.001). In Cox regression analysis, AKI stage was independently associated with primary endpoint, with an adjusted hazard ratio (HR) of 3.8(95% CI 2.1-6.7, P < 0.001). CONCLUSION: Severity of AKI based on KDIGO AKI category was associated with progression to ESRD in DPLN patients. Analytical data also confirmed the good discriminative power of the KDIGO AKI classification system for predicting long-term prognosis of DPLN patients with AKI.

Antimicrobial therapy with aminoglycoside or meropenem in the intensive care unit for hospital associated infections and risk factors for acute kidney injury.

There have historically been concerns of acute kidney injury (AKI) with the use of aminoglycosides. The present study aimed to compare the AKI incidence and mortality rate between critically ill patients treated with aminoglycoside or meropenem in the intensive care unit setting using a propensity score matching approach. This cross-sectional study was conducted at two university hospitals from January 2011 to October 2017. Clinical and laboratorial data were evaluated to exclude potential confounders and to calculate the Charlson index. AKI was classified according to the Acute Kidney Injury Network criteria. All tests were two-tailed, and a p value ≤ 0.05 was considered significant in the univariate and multivariate analyses. We included 494 patients, 95 and 399 of whom used meropenem and aminoglycoside, respectively. Patients in the subgroup that used meropenem were matched with controls (aminoglycoside). Among the 494 patients, 120 developed any grade of AKI (24.2%). After propensity score matching, there were no significant differences in AKI incidence and mortality rate between the aminoglycoside and meropenem groups (p = 0.324 and 0.464, respectively). Patients on the aminoglycoside regimen neither presented a higher AKI incidence nor mortality rate when compared with those on the meropenem regimen. Aminoglycosides may be a safe option for the treatment of critically ill patients on carbapenem sparing antimicrobial stewardship programs.